Coverage Report

Created: 2026-04-09 11:07

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/exec/pipeline/pipeline_fragment_context.cpp
Line
Count
Source
1
// Licensed to the Apache Software Foundation (ASF) under one
2
// or more contributor license agreements.  See the NOTICE file
3
// distributed with this work for additional information
4
// regarding copyright ownership.  The ASF licenses this file
5
// to you under the Apache License, Version 2.0 (the
6
// "License"); you may not use this file except in compliance
7
// with the License.  You may obtain a copy of the License at
8
//
9
//   http://www.apache.org/licenses/LICENSE-2.0
10
//
11
// Unless required by applicable law or agreed to in writing,
12
// software distributed under the License is distributed on an
13
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14
// KIND, either express or implied.  See the License for the
15
// specific language governing permissions and limitations
16
// under the License.
17
18
#include "exec/pipeline/pipeline_fragment_context.h"
19
20
#include <gen_cpp/DataSinks_types.h>
21
#include <gen_cpp/PaloInternalService_types.h>
22
#include <gen_cpp/PlanNodes_types.h>
23
#include <pthread.h>
24
25
#include <algorithm>
26
#include <cstdlib>
27
// IWYU pragma: no_include <bits/chrono.h>
28
#include <fmt/format.h>
29
30
#include <chrono> // IWYU pragma: keep
31
#include <map>
32
#include <memory>
33
#include <ostream>
34
#include <utility>
35
36
#include "cloud/config.h"
37
#include "common/cast_set.h"
38
#include "common/config.h"
39
#include "common/exception.h"
40
#include "common/logging.h"
41
#include "common/status.h"
42
#include "exec/exchange/local_exchange_sink_operator.h"
43
#include "exec/exchange/local_exchange_source_operator.h"
44
#include "exec/exchange/local_exchanger.h"
45
#include "exec/exchange/vdata_stream_mgr.h"
46
#include "exec/operator/aggregation_sink_operator.h"
47
#include "exec/operator/aggregation_source_operator.h"
48
#include "exec/operator/analytic_sink_operator.h"
49
#include "exec/operator/analytic_source_operator.h"
50
#include "exec/operator/assert_num_rows_operator.h"
51
#include "exec/operator/blackhole_sink_operator.h"
52
#include "exec/operator/cache_sink_operator.h"
53
#include "exec/operator/cache_source_operator.h"
54
#include "exec/operator/datagen_operator.h"
55
#include "exec/operator/dict_sink_operator.h"
56
#include "exec/operator/distinct_streaming_aggregation_operator.h"
57
#include "exec/operator/empty_set_operator.h"
58
#include "exec/operator/es_scan_operator.h"
59
#include "exec/operator/exchange_sink_operator.h"
60
#include "exec/operator/exchange_source_operator.h"
61
#include "exec/operator/file_scan_operator.h"
62
#include "exec/operator/group_commit_block_sink_operator.h"
63
#include "exec/operator/group_commit_scan_operator.h"
64
#include "exec/operator/hashjoin_build_sink.h"
65
#include "exec/operator/hashjoin_probe_operator.h"
66
#include "exec/operator/hive_table_sink_operator.h"
67
#include "exec/operator/iceberg_delete_sink_operator.h"
68
#include "exec/operator/iceberg_merge_sink_operator.h"
69
#include "exec/operator/iceberg_table_sink_operator.h"
70
#include "exec/operator/jdbc_scan_operator.h"
71
#include "exec/operator/jdbc_table_sink_operator.h"
72
#include "exec/operator/local_merge_sort_source_operator.h"
73
#include "exec/operator/materialization_opertor.h"
74
#include "exec/operator/maxcompute_table_sink_operator.h"
75
#include "exec/operator/memory_scratch_sink_operator.h"
76
#include "exec/operator/meta_scan_operator.h"
77
#include "exec/operator/multi_cast_data_stream_sink.h"
78
#include "exec/operator/multi_cast_data_stream_source.h"
79
#include "exec/operator/nested_loop_join_build_operator.h"
80
#include "exec/operator/nested_loop_join_probe_operator.h"
81
#include "exec/operator/olap_scan_operator.h"
82
#include "exec/operator/olap_table_sink_operator.h"
83
#include "exec/operator/olap_table_sink_v2_operator.h"
84
#include "exec/operator/partition_sort_sink_operator.h"
85
#include "exec/operator/partition_sort_source_operator.h"
86
#include "exec/operator/partitioned_aggregation_sink_operator.h"
87
#include "exec/operator/partitioned_aggregation_source_operator.h"
88
#include "exec/operator/partitioned_hash_join_probe_operator.h"
89
#include "exec/operator/partitioned_hash_join_sink_operator.h"
90
#include "exec/operator/rec_cte_anchor_sink_operator.h"
91
#include "exec/operator/rec_cte_scan_operator.h"
92
#include "exec/operator/rec_cte_sink_operator.h"
93
#include "exec/operator/rec_cte_source_operator.h"
94
#include "exec/operator/repeat_operator.h"
95
#include "exec/operator/result_file_sink_operator.h"
96
#include "exec/operator/result_sink_operator.h"
97
#include "exec/operator/schema_scan_operator.h"
98
#include "exec/operator/select_operator.h"
99
#include "exec/operator/set_probe_sink_operator.h"
100
#include "exec/operator/set_sink_operator.h"
101
#include "exec/operator/set_source_operator.h"
102
#include "exec/operator/sort_sink_operator.h"
103
#include "exec/operator/sort_source_operator.h"
104
#include "exec/operator/spill_iceberg_table_sink_operator.h"
105
#include "exec/operator/spill_sort_sink_operator.h"
106
#include "exec/operator/spill_sort_source_operator.h"
107
#include "exec/operator/streaming_aggregation_operator.h"
108
#include "exec/operator/table_function_operator.h"
109
#include "exec/operator/tvf_table_sink_operator.h"
110
#include "exec/operator/union_sink_operator.h"
111
#include "exec/operator/union_source_operator.h"
112
#include "exec/pipeline/dependency.h"
113
#include "exec/pipeline/pipeline_task.h"
114
#include "exec/pipeline/task_scheduler.h"
115
#include "exec/runtime_filter/runtime_filter_mgr.h"
116
#include "exec/sort/topn_sorter.h"
117
#include "exec/spill/spill_file.h"
118
#include "io/fs/stream_load_pipe.h"
119
#include "load/stream_load/new_load_stream_mgr.h"
120
#include "runtime/exec_env.h"
121
#include "runtime/fragment_mgr.h"
122
#include "runtime/result_buffer_mgr.h"
123
#include "runtime/runtime_state.h"
124
#include "runtime/thread_context.h"
125
#include "util/countdown_latch.h"
126
#include "util/debug_util.h"
127
#include "util/uid_util.h"
128
129
namespace doris {
130
#include "common/compile_check_begin.h"
131
PipelineFragmentContext::PipelineFragmentContext(
132
        TUniqueId query_id, const TPipelineFragmentParams& request,
133
        std::shared_ptr<QueryContext> query_ctx, ExecEnv* exec_env,
134
        const std::function<void(RuntimeState*, Status*)>& call_back,
135
        report_status_callback report_status_cb)
136
430k
        : _query_id(std::move(query_id)),
137
430k
          _fragment_id(request.fragment_id),
138
430k
          _exec_env(exec_env),
139
430k
          _query_ctx(std::move(query_ctx)),
140
430k
          _call_back(call_back),
141
430k
          _is_report_on_cancel(true),
142
430k
          _report_status_cb(std::move(report_status_cb)),
143
430k
          _params(request),
144
430k
          _parallel_instances(_params.__isset.parallel_instances ? _params.parallel_instances : 0),
145
430k
          _need_notify_close(request.__isset.need_notify_close ? request.need_notify_close
146
430k
                                                               : false) {
147
430k
    _fragment_watcher.start();
148
430k
}
149
150
431k
PipelineFragmentContext::~PipelineFragmentContext() {
151
431k
    LOG_INFO("PipelineFragmentContext::~PipelineFragmentContext")
152
431k
            .tag("query_id", print_id(_query_id))
153
431k
            .tag("fragment_id", _fragment_id);
154
431k
    _release_resource();
155
431k
    {
156
        // The memory released by the query end is recorded in the query mem tracker.
157
431k
        SCOPED_SWITCH_THREAD_MEM_TRACKER_LIMITER(_query_ctx->query_mem_tracker());
158
431k
        _runtime_state.reset();
159
431k
        _query_ctx.reset();
160
431k
    }
161
431k
}
162
163
60
bool PipelineFragmentContext::is_timeout(timespec now) const {
164
60
    if (_timeout <= 0) {
165
0
        return false;
166
0
    }
167
60
    return _fragment_watcher.elapsed_time_seconds(now) > _timeout;
168
60
}
169
170
// notify_close() transitions the PFC from "waiting for external close notification" to
171
// "self-managed close". For recursive CTE fragments, the old PFC is kept alive until
172
// the rerun_fragment(wait_for_destroy) RPC calls this to trigger shutdown.
173
// Returns true if all tasks have already closed (i.e., the PFC can be safely destroyed).
174
10.4k
bool PipelineFragmentContext::notify_close() {
175
10.4k
    bool all_closed = false;
176
10.4k
    bool need_remove = false;
177
10.4k
    {
178
10.4k
        std::lock_guard<std::mutex> l(_task_mutex);
179
10.4k
        if (_closed_tasks >= _total_tasks) {
180
3.44k
            if (_need_notify_close) {
181
                // Fragment was cancelled and waiting for notify to close.
182
                // Record that we need to remove from fragment mgr, but do it
183
                // after releasing _task_mutex to avoid ABBA deadlock with
184
                // dump_pipeline_tasks() (which acquires _pipeline_map lock
185
                // first, then _task_mutex via debug_string()).
186
3.38k
                need_remove = true;
187
3.38k
            }
188
3.44k
            all_closed = true;
189
3.44k
        }
190
        // make fragment release by self after cancel
191
10.4k
        _need_notify_close = false;
192
10.4k
    }
193
10.4k
    if (need_remove) {
194
3.38k
        _exec_env->fragment_mgr()->remove_pipeline_context({_query_id, _fragment_id});
195
3.38k
    }
196
10.4k
    return all_closed;
197
10.4k
}
198
199
// Must not add lock in this method. Because it will call query ctx cancel. And
200
// QueryCtx cancel will call fragment ctx cancel. And Also Fragment ctx's running
201
// Method like exchange sink buffer will call query ctx cancel. If we add lock here
202
// There maybe dead lock.
203
6.92k
void PipelineFragmentContext::cancel(const Status reason) {
204
6.92k
    LOG_INFO("PipelineFragmentContext::cancel")
205
6.92k
            .tag("query_id", print_id(_query_id))
206
6.92k
            .tag("fragment_id", _fragment_id)
207
6.92k
            .tag("reason", reason.to_string());
208
6.92k
    if (notify_close()) {
209
81
        return;
210
81
    }
211
    // Timeout is a special error code, we need print current stack to debug timeout issue.
212
6.84k
    if (reason.is<ErrorCode::TIMEOUT>()) {
213
1
        auto dbg_str = fmt::format("PipelineFragmentContext is cancelled due to timeout:\n{}",
214
1
                                   debug_string());
215
1
        LOG_LONG_STRING(WARNING, dbg_str);
216
1
    }
217
218
    // `ILLEGAL_STATE` means queries this fragment belongs to was not found in FE (maybe finished)
219
6.84k
    if (reason.is<ErrorCode::ILLEGAL_STATE>()) {
220
0
        LOG_WARNING("PipelineFragmentContext is cancelled due to illegal state : {}",
221
0
                    debug_string());
222
0
    }
223
224
6.84k
    if (reason.is<ErrorCode::MEM_LIMIT_EXCEEDED>() || reason.is<ErrorCode::MEM_ALLOC_FAILED>()) {
225
0
        print_profile("cancel pipeline, reason: " + reason.to_string());
226
0
    }
227
228
6.84k
    if (auto error_url = get_load_error_url(); !error_url.empty()) {
229
23
        _query_ctx->set_load_error_url(error_url);
230
23
    }
231
232
6.84k
    if (auto first_error_msg = get_first_error_msg(); !first_error_msg.empty()) {
233
23
        _query_ctx->set_first_error_msg(first_error_msg);
234
23
    }
235
236
6.84k
    _query_ctx->cancel(reason, _fragment_id);
237
6.84k
    if (reason.is<ErrorCode::LIMIT_REACH>()) {
238
491
        _is_report_on_cancel = false;
239
6.35k
    } else {
240
41.0k
        for (auto& id : _fragment_instance_ids) {
241
41.0k
            LOG(WARNING) << "PipelineFragmentContext cancel instance: " << print_id(id);
242
41.0k
        }
243
6.35k
    }
244
    // Get pipe from new load stream manager and send cancel to it or the fragment may hang to wait read from pipe
245
    // For stream load the fragment's query_id == load id, it is set in FE.
246
6.84k
    auto stream_load_ctx = _exec_env->new_load_stream_mgr()->get(_query_id);
247
6.84k
    if (stream_load_ctx != nullptr) {
248
31
        stream_load_ctx->pipe->cancel(reason.to_string());
249
        // Set error URL here because after pipe is cancelled, stream load execution may return early.
250
        // We need to set the error URL at this point to ensure error information is properly
251
        // propagated to the client.
252
31
        stream_load_ctx->error_url = get_load_error_url();
253
31
        stream_load_ctx->first_error_msg = get_first_error_msg();
254
31
    }
255
256
43.0k
    for (auto& tasks : _tasks) {
257
94.5k
        for (auto& task : tasks) {
258
94.5k
            task.first->unblock_all_dependencies();
259
94.5k
        }
260
43.0k
    }
261
6.84k
}
262
263
671k
PipelinePtr PipelineFragmentContext::add_pipeline(PipelinePtr parent, int idx) {
264
671k
    PipelineId id = _next_pipeline_id++;
265
671k
    auto pipeline = std::make_shared<Pipeline>(
266
671k
            id, parent ? std::min(parent->num_tasks(), _num_instances) : _num_instances,
267
671k
            parent ? parent->num_tasks() : _num_instances);
268
671k
    if (idx >= 0) {
269
111k
        _pipelines.insert(_pipelines.begin() + idx, pipeline);
270
559k
    } else {
271
559k
        _pipelines.emplace_back(pipeline);
272
559k
    }
273
671k
    if (parent) {
274
235k
        parent->set_children(pipeline);
275
235k
    }
276
671k
    return pipeline;
277
671k
}
278
279
430k
Status PipelineFragmentContext::_build_and_prepare_full_pipeline(ThreadPool* thread_pool) {
280
430k
    {
281
430k
        SCOPED_TIMER(_build_pipelines_timer);
282
        // 2. Build pipelines with operators in this fragment.
283
430k
        auto root_pipeline = add_pipeline();
284
430k
        RETURN_IF_ERROR(_build_pipelines(_runtime_state->obj_pool(), *_query_ctx->desc_tbl,
285
430k
                                         &_root_op, root_pipeline));
286
287
        // 3. Create sink operator
288
430k
        if (!_params.fragment.__isset.output_sink) {
289
0
            return Status::InternalError("No output sink in this fragment!");
290
0
        }
291
430k
        RETURN_IF_ERROR(_create_data_sink(_runtime_state->obj_pool(), _params.fragment.output_sink,
292
430k
                                          _params.fragment.output_exprs, _params,
293
430k
                                          root_pipeline->output_row_desc(), _runtime_state.get(),
294
430k
                                          *_desc_tbl, root_pipeline->id()));
295
430k
        RETURN_IF_ERROR(_sink->init(_params.fragment.output_sink));
296
430k
        RETURN_IF_ERROR(root_pipeline->set_sink(_sink));
297
298
559k
        for (PipelinePtr& pipeline : _pipelines) {
299
18.4E
            DCHECK(pipeline->sink() != nullptr) << pipeline->operators().size();
300
559k
            RETURN_IF_ERROR(pipeline->sink()->set_child(pipeline->operators().back()));
301
559k
        }
302
430k
    }
303
    // 4. Build local exchanger
304
430k
    if (_runtime_state->enable_local_shuffle()) {
305
427k
        SCOPED_TIMER(_plan_local_exchanger_timer);
306
427k
        RETURN_IF_ERROR(_plan_local_exchange(_params.num_buckets,
307
427k
                                             _params.bucket_seq_to_instance_idx,
308
427k
                                             _params.shuffle_idx_to_instance_idx));
309
427k
    }
310
311
    // 5. Initialize global states in pipelines.
312
672k
    for (PipelinePtr& pipeline : _pipelines) {
313
672k
        SCOPED_TIMER(_prepare_all_pipelines_timer);
314
672k
        pipeline->children().clear();
315
672k
        RETURN_IF_ERROR(pipeline->prepare(_runtime_state.get()));
316
672k
    }
317
318
428k
    {
319
428k
        SCOPED_TIMER(_build_tasks_timer);
320
        // 6. Build pipeline tasks and initialize local state.
321
428k
        RETURN_IF_ERROR(_build_pipeline_tasks(thread_pool));
322
428k
    }
323
324
428k
    return Status::OK();
325
428k
}
326
327
430k
Status PipelineFragmentContext::prepare(ThreadPool* thread_pool) {
328
430k
    if (_prepared) {
329
0
        return Status::InternalError("Already prepared");
330
0
    }
331
430k
    if (_params.__isset.query_options && _params.query_options.__isset.execution_timeout) {
332
430k
        _timeout = _params.query_options.execution_timeout;
333
430k
    }
334
335
430k
    _fragment_level_profile = std::make_unique<RuntimeProfile>("PipelineContext");
336
430k
    _prepare_timer = ADD_TIMER(_fragment_level_profile, "PrepareTime");
337
430k
    SCOPED_TIMER(_prepare_timer);
338
430k
    _build_pipelines_timer = ADD_TIMER(_fragment_level_profile, "BuildPipelinesTime");
339
430k
    _init_context_timer = ADD_TIMER(_fragment_level_profile, "InitContextTime");
340
430k
    _plan_local_exchanger_timer = ADD_TIMER(_fragment_level_profile, "PlanLocalLocalExchangerTime");
341
430k
    _build_tasks_timer = ADD_TIMER(_fragment_level_profile, "BuildTasksTime");
342
430k
    _prepare_all_pipelines_timer = ADD_TIMER(_fragment_level_profile, "PrepareAllPipelinesTime");
343
430k
    {
344
430k
        SCOPED_TIMER(_init_context_timer);
345
430k
        cast_set(_num_instances, _params.local_params.size());
346
430k
        _total_instances =
347
430k
                _params.__isset.total_instances ? _params.total_instances : _num_instances;
348
349
430k
        auto* fragment_context = this;
350
351
430k
        if (_params.query_options.__isset.is_report_success) {
352
428k
            fragment_context->set_is_report_success(_params.query_options.is_report_success);
353
428k
        }
354
355
        // 1. Set up the global runtime state.
356
430k
        _runtime_state = RuntimeState::create_unique(
357
430k
                _params.query_id, _params.fragment_id, _params.query_options,
358
430k
                _query_ctx->query_globals, _exec_env, _query_ctx.get());
359
430k
        _runtime_state->set_task_execution_context(shared_from_this());
360
430k
        SCOPED_SWITCH_THREAD_MEM_TRACKER_LIMITER(_runtime_state->query_mem_tracker());
361
430k
        if (_params.__isset.backend_id) {
362
428k
            _runtime_state->set_backend_id(_params.backend_id);
363
428k
        }
364
430k
        if (_params.__isset.import_label) {
365
239
            _runtime_state->set_import_label(_params.import_label);
366
239
        }
367
430k
        if (_params.__isset.db_name) {
368
191
            _runtime_state->set_db_name(_params.db_name);
369
191
        }
370
430k
        if (_params.__isset.load_job_id) {
371
0
            _runtime_state->set_load_job_id(_params.load_job_id);
372
0
        }
373
374
430k
        if (_params.is_simplified_param) {
375
145k
            _desc_tbl = _query_ctx->desc_tbl;
376
284k
        } else {
377
284k
            DCHECK(_params.__isset.desc_tbl);
378
284k
            RETURN_IF_ERROR(DescriptorTbl::create(_runtime_state->obj_pool(), _params.desc_tbl,
379
284k
                                                  &_desc_tbl));
380
284k
        }
381
430k
        _runtime_state->set_desc_tbl(_desc_tbl);
382
430k
        _runtime_state->set_num_per_fragment_instances(_params.num_senders);
383
430k
        _runtime_state->set_load_stream_per_node(_params.load_stream_per_node);
384
430k
        _runtime_state->set_total_load_streams(_params.total_load_streams);
385
430k
        _runtime_state->set_num_local_sink(_params.num_local_sink);
386
387
        // init fragment_instance_ids
388
430k
        const auto target_size = _params.local_params.size();
389
430k
        _fragment_instance_ids.resize(target_size);
390
1.55M
        for (size_t i = 0; i < _params.local_params.size(); i++) {
391
1.12M
            auto fragment_instance_id = _params.local_params[i].fragment_instance_id;
392
1.12M
            _fragment_instance_ids[i] = fragment_instance_id;
393
1.12M
        }
394
430k
    }
395
396
430k
    RETURN_IF_ERROR(_build_and_prepare_full_pipeline(thread_pool));
397
398
429k
    _init_next_report_time();
399
400
429k
    _prepared = true;
401
429k
    return Status::OK();
402
430k
}
403
404
Status PipelineFragmentContext::_build_pipeline_tasks_for_instance(
405
        int instance_idx,
406
1.12M
        const std::vector<std::shared_ptr<RuntimeProfile>>& pipeline_id_to_profile) {
407
1.12M
    const auto& local_params = _params.local_params[instance_idx];
408
1.12M
    auto fragment_instance_id = local_params.fragment_instance_id;
409
1.12M
    auto runtime_filter_mgr = std::make_unique<RuntimeFilterMgr>(false);
410
1.12M
    std::map<PipelineId, PipelineTask*> pipeline_id_to_task;
411
1.12M
    auto get_shared_state = [&](PipelinePtr pipeline)
412
1.12M
            -> std::map<int, std::pair<std::shared_ptr<BasicSharedState>,
413
1.92M
                                       std::vector<std::shared_ptr<Dependency>>>> {
414
1.92M
        std::map<int, std::pair<std::shared_ptr<BasicSharedState>,
415
1.92M
                                std::vector<std::shared_ptr<Dependency>>>>
416
1.92M
                shared_state_map;
417
2.47M
        for (auto& op : pipeline->operators()) {
418
2.47M
            auto source_id = op->operator_id();
419
2.47M
            if (auto iter = _op_id_to_shared_state.find(source_id);
420
2.47M
                iter != _op_id_to_shared_state.end()) {
421
715k
                shared_state_map.insert({source_id, iter->second});
422
715k
            }
423
2.47M
        }
424
1.92M
        for (auto sink_to_source_id : pipeline->sink()->dests_id()) {
425
1.92M
            if (auto iter = _op_id_to_shared_state.find(sink_to_source_id);
426
1.92M
                iter != _op_id_to_shared_state.end()) {
427
306k
                shared_state_map.insert({sink_to_source_id, iter->second});
428
306k
            }
429
1.92M
        }
430
1.92M
        return shared_state_map;
431
1.92M
    };
432
433
3.45M
    for (size_t pip_idx = 0; pip_idx < _pipelines.size(); pip_idx++) {
434
2.32M
        auto& pipeline = _pipelines[pip_idx];
435
2.32M
        if (pipeline->num_tasks() > 1 || instance_idx == 0) {
436
1.91M
            auto task_runtime_state = RuntimeState::create_unique(
437
1.91M
                    local_params.fragment_instance_id, _params.query_id, _params.fragment_id,
438
1.91M
                    _params.query_options, _query_ctx->query_globals, _exec_env, _query_ctx.get());
439
1.91M
            {
440
                // Initialize runtime state for this task
441
1.91M
                task_runtime_state->set_query_mem_tracker(_query_ctx->query_mem_tracker());
442
443
1.91M
                task_runtime_state->set_task_execution_context(shared_from_this());
444
1.91M
                task_runtime_state->set_be_number(local_params.backend_num);
445
446
1.92M
                if (_params.__isset.backend_id) {
447
1.92M
                    task_runtime_state->set_backend_id(_params.backend_id);
448
1.92M
                }
449
1.91M
                if (_params.__isset.import_label) {
450
240
                    task_runtime_state->set_import_label(_params.import_label);
451
240
                }
452
1.91M
                if (_params.__isset.db_name) {
453
192
                    task_runtime_state->set_db_name(_params.db_name);
454
192
                }
455
1.91M
                if (_params.__isset.load_job_id) {
456
0
                    task_runtime_state->set_load_job_id(_params.load_job_id);
457
0
                }
458
1.91M
                if (_params.__isset.wal_id) {
459
114
                    task_runtime_state->set_wal_id(_params.wal_id);
460
114
                }
461
1.91M
                if (_params.__isset.content_length) {
462
31
                    task_runtime_state->set_content_length(_params.content_length);
463
31
                }
464
465
1.91M
                task_runtime_state->set_desc_tbl(_desc_tbl);
466
1.91M
                task_runtime_state->set_per_fragment_instance_idx(local_params.sender_id);
467
1.91M
                task_runtime_state->set_num_per_fragment_instances(_params.num_senders);
468
1.91M
                task_runtime_state->resize_op_id_to_local_state(max_operator_id());
469
1.91M
                task_runtime_state->set_max_operator_id(max_operator_id());
470
1.91M
                task_runtime_state->set_load_stream_per_node(_params.load_stream_per_node);
471
1.91M
                task_runtime_state->set_total_load_streams(_params.total_load_streams);
472
1.91M
                task_runtime_state->set_num_local_sink(_params.num_local_sink);
473
474
1.91M
                task_runtime_state->set_runtime_filter_mgr(runtime_filter_mgr.get());
475
1.91M
            }
476
1.91M
            auto cur_task_id = _total_tasks++;
477
1.91M
            task_runtime_state->set_task_id(cur_task_id);
478
1.91M
            task_runtime_state->set_task_num(pipeline->num_tasks());
479
1.91M
            auto task = std::make_shared<PipelineTask>(
480
1.91M
                    pipeline, cur_task_id, task_runtime_state.get(),
481
1.91M
                    std::dynamic_pointer_cast<PipelineFragmentContext>(shared_from_this()),
482
1.91M
                    pipeline_id_to_profile[pip_idx].get(), get_shared_state(pipeline),
483
1.91M
                    instance_idx);
484
1.91M
            pipeline->incr_created_tasks(instance_idx, task.get());
485
1.91M
            pipeline_id_to_task.insert({pipeline->id(), task.get()});
486
1.91M
            _tasks[instance_idx].emplace_back(
487
1.91M
                    std::pair<std::shared_ptr<PipelineTask>, std::unique_ptr<RuntimeState>> {
488
1.91M
                            std::move(task), std::move(task_runtime_state)});
489
1.91M
        }
490
2.32M
    }
491
492
    /**
493
         * Build DAG for pipeline tasks.
494
         * For example, we have
495
         *
496
         *   ExchangeSink (Pipeline1)     JoinBuildSink (Pipeline2)
497
         *            \                      /
498
         *          JoinProbeOperator1 (Pipeline1)    JoinBuildSink (Pipeline3)
499
         *                 \                          /
500
         *               JoinProbeOperator2 (Pipeline1)
501
         *
502
         * In this fragment, we have three pipelines and pipeline 1 depends on pipeline 2 and pipeline 3.
503
         * To build this DAG, `_dag` manage dependencies between pipelines by pipeline ID and
504
         * `pipeline_id_to_task` is used to find the task by a unique pipeline ID.
505
         *
506
         * Finally, we have two upstream dependencies in Pipeline1 corresponding to JoinProbeOperator1
507
         * and JoinProbeOperator2.
508
         */
509
2.32M
    for (auto& _pipeline : _pipelines) {
510
2.32M
        if (pipeline_id_to_task.contains(_pipeline->id())) {
511
1.91M
            auto* task = pipeline_id_to_task[_pipeline->id()];
512
1.91M
            DCHECK(task != nullptr);
513
514
            // If this task has upstream dependency, then inject it into this task.
515
1.91M
            if (_dag.contains(_pipeline->id())) {
516
1.19M
                auto& deps = _dag[_pipeline->id()];
517
1.89M
                for (auto& dep : deps) {
518
1.89M
                    if (pipeline_id_to_task.contains(dep)) {
519
1.07M
                        auto ss = pipeline_id_to_task[dep]->get_sink_shared_state();
520
1.07M
                        if (ss) {
521
476k
                            task->inject_shared_state(ss);
522
599k
                        } else {
523
599k
                            pipeline_id_to_task[dep]->inject_shared_state(
524
599k
                                    task->get_source_shared_state());
525
599k
                        }
526
1.07M
                    }
527
1.89M
                }
528
1.19M
            }
529
1.91M
        }
530
2.32M
    }
531
3.45M
    for (size_t pip_idx = 0; pip_idx < _pipelines.size(); pip_idx++) {
532
2.32M
        if (pipeline_id_to_task.contains(_pipelines[pip_idx]->id())) {
533
1.91M
            auto* task = pipeline_id_to_task[_pipelines[pip_idx]->id()];
534
1.91M
            DCHECK(pipeline_id_to_profile[pip_idx]);
535
1.91M
            std::vector<TScanRangeParams> scan_ranges;
536
1.91M
            auto node_id = _pipelines[pip_idx]->operators().front()->node_id();
537
1.91M
            if (local_params.per_node_scan_ranges.contains(node_id)) {
538
347k
                scan_ranges = local_params.per_node_scan_ranges.find(node_id)->second;
539
347k
            }
540
1.91M
            RETURN_IF_ERROR_OR_CATCH_EXCEPTION(task->prepare(scan_ranges, local_params.sender_id,
541
1.91M
                                                             _params.fragment.output_sink));
542
1.91M
        }
543
2.32M
    }
544
1.13M
    {
545
1.13M
        std::lock_guard<std::mutex> l(_state_map_lock);
546
1.13M
        _runtime_filter_mgr_map[instance_idx] = std::move(runtime_filter_mgr);
547
1.13M
    }
548
1.13M
    return Status::OK();
549
1.12M
}
550
551
429k
Status PipelineFragmentContext::_build_pipeline_tasks(ThreadPool* thread_pool) {
552
429k
    _total_tasks = 0;
553
429k
    _closed_tasks = 0;
554
429k
    const auto target_size = _params.local_params.size();
555
429k
    _tasks.resize(target_size);
556
429k
    _runtime_filter_mgr_map.resize(target_size);
557
1.10M
    for (size_t pip_idx = 0; pip_idx < _pipelines.size(); pip_idx++) {
558
671k
        _pip_id_to_pipeline[_pipelines[pip_idx]->id()] = _pipelines[pip_idx].get();
559
671k
    }
560
429k
    auto pipeline_id_to_profile = _runtime_state->build_pipeline_profile(_pipelines.size());
561
562
429k
    if (target_size > 1 &&
563
429k
        (_runtime_state->query_options().__isset.parallel_prepare_threshold &&
564
133k
         target_size > _runtime_state->query_options().parallel_prepare_threshold)) {
565
        // If instances parallelism is big enough ( > parallel_prepare_threshold), we will prepare all tasks by multi-threads
566
19.7k
        std::vector<Status> prepare_status(target_size);
567
19.7k
        int submitted_tasks = 0;
568
19.7k
        Status submit_status;
569
19.7k
        CountDownLatch latch((int)target_size);
570
247k
        for (int i = 0; i < target_size; i++) {
571
227k
            submit_status = thread_pool->submit_func([&, i]() {
572
227k
                SCOPED_ATTACH_TASK(_query_ctx.get());
573
227k
                prepare_status[i] = _build_pipeline_tasks_for_instance(i, pipeline_id_to_profile);
574
227k
                latch.count_down();
575
227k
            });
576
227k
            if (LIKELY(submit_status.ok())) {
577
227k
                submitted_tasks++;
578
18.4E
            } else {
579
18.4E
                break;
580
18.4E
            }
581
227k
        }
582
19.7k
        latch.arrive_and_wait(target_size - submitted_tasks);
583
19.7k
        if (UNLIKELY(!submit_status.ok())) {
584
0
            return submit_status;
585
0
        }
586
247k
        for (int i = 0; i < submitted_tasks; i++) {
587
227k
            if (!prepare_status[i].ok()) {
588
0
                return prepare_status[i];
589
0
            }
590
227k
        }
591
410k
    } else {
592
1.31M
        for (int i = 0; i < target_size; i++) {
593
902k
            RETURN_IF_ERROR(_build_pipeline_tasks_for_instance(i, pipeline_id_to_profile));
594
902k
        }
595
410k
    }
596
429k
    _pipeline_parent_map.clear();
597
429k
    _op_id_to_shared_state.clear();
598
599
429k
    return Status::OK();
600
429k
}
601
602
428k
void PipelineFragmentContext::_init_next_report_time() {
603
428k
    auto interval_s = config::pipeline_status_report_interval;
604
428k
    if (_is_report_success && interval_s > 0 && _timeout > interval_s) {
605
40.2k
        VLOG_FILE << "enable period report: fragment id=" << _fragment_id;
606
40.2k
        uint64_t report_fragment_offset = (uint64_t)(rand() % interval_s) * NANOS_PER_SEC;
607
        // We don't want to wait longer than it takes to run the entire fragment.
608
40.2k
        _previous_report_time =
609
40.2k
                MonotonicNanos() + report_fragment_offset - (uint64_t)(interval_s)*NANOS_PER_SEC;
610
40.2k
        _disable_period_report = false;
611
40.2k
    }
612
428k
}
613
614
4.68k
void PipelineFragmentContext::refresh_next_report_time() {
615
4.68k
    auto disable = _disable_period_report.load(std::memory_order_acquire);
616
4.68k
    DCHECK(disable == true);
617
4.68k
    _previous_report_time.store(MonotonicNanos(), std::memory_order_release);
618
4.68k
    _disable_period_report.compare_exchange_strong(disable, false);
619
4.68k
}
620
621
6.92M
void PipelineFragmentContext::trigger_report_if_necessary() {
622
6.92M
    if (!_is_report_success) {
623
6.40M
        return;
624
6.40M
    }
625
518k
    auto disable = _disable_period_report.load(std::memory_order_acquire);
626
518k
    if (disable) {
627
8.91k
        return;
628
8.91k
    }
629
509k
    int32_t interval_s = config::pipeline_status_report_interval;
630
509k
    if (interval_s <= 0) {
631
0
        LOG(WARNING) << "config::status_report_interval is equal to or less than zero, do not "
632
0
                        "trigger "
633
0
                        "report.";
634
0
    }
635
509k
    uint64_t next_report_time = _previous_report_time.load(std::memory_order_acquire) +
636
509k
                                (uint64_t)(interval_s)*NANOS_PER_SEC;
637
509k
    if (MonotonicNanos() > next_report_time) {
638
4.70k
        if (!_disable_period_report.compare_exchange_strong(disable, true,
639
4.70k
                                                            std::memory_order_acq_rel)) {
640
17
            return;
641
17
        }
642
4.68k
        if (VLOG_FILE_IS_ON) {
643
0
            VLOG_FILE << "Reporting "
644
0
                      << "profile for query_id " << print_id(_query_id)
645
0
                      << ", fragment id: " << _fragment_id;
646
647
0
            std::stringstream ss;
648
0
            _runtime_state->runtime_profile()->compute_time_in_profile();
649
0
            _runtime_state->runtime_profile()->pretty_print(&ss);
650
0
            if (_runtime_state->load_channel_profile()) {
651
0
                _runtime_state->load_channel_profile()->pretty_print(&ss);
652
0
            }
653
654
0
            VLOG_FILE << "Query " << print_id(get_query_id()) << " fragment " << get_fragment_id()
655
0
                      << " profile:\n"
656
0
                      << ss.str();
657
0
        }
658
4.68k
        auto st = send_report(false);
659
4.68k
        if (!st.ok()) {
660
0
            disable = true;
661
0
            _disable_period_report.compare_exchange_strong(disable, false,
662
0
                                                           std::memory_order_acq_rel);
663
0
        }
664
4.68k
    }
665
509k
}
666
667
Status PipelineFragmentContext::_build_pipelines(ObjectPool* pool, const DescriptorTbl& descs,
668
426k
                                                 OperatorPtr* root, PipelinePtr cur_pipe) {
669
426k
    if (_params.fragment.plan.nodes.empty()) {
670
0
        throw Exception(ErrorCode::INTERNAL_ERROR, "Invalid plan which has no plan node!");
671
0
    }
672
673
426k
    int node_idx = 0;
674
675
426k
    RETURN_IF_ERROR(_create_tree_helper(pool, _params.fragment.plan.nodes, descs, nullptr,
676
426k
                                        &node_idx, root, cur_pipe, 0, false, false));
677
678
426k
    if (node_idx + 1 != _params.fragment.plan.nodes.size()) {
679
0
        return Status::InternalError(
680
0
                "Plan tree only partially reconstructed. Not all thrift nodes were used.");
681
0
    }
682
426k
    return Status::OK();
683
426k
}
684
685
Status PipelineFragmentContext::_create_tree_helper(
686
        ObjectPool* pool, const std::vector<TPlanNode>& tnodes, const DescriptorTbl& descs,
687
        OperatorPtr parent, int* node_idx, OperatorPtr* root, PipelinePtr& cur_pipe, int child_idx,
688
663k
        const bool followed_by_shuffled_operator, const bool require_bucket_distribution) {
689
    // propagate error case
690
663k
    if (*node_idx >= tnodes.size()) {
691
0
        return Status::InternalError(
692
0
                "Failed to reconstruct plan tree from thrift. Node id: {}, number of nodes: {}",
693
0
                *node_idx, tnodes.size());
694
0
    }
695
663k
    const TPlanNode& tnode = tnodes[*node_idx];
696
697
663k
    int num_children = tnodes[*node_idx].num_children;
698
663k
    bool current_followed_by_shuffled_operator = followed_by_shuffled_operator;
699
663k
    bool current_require_bucket_distribution = require_bucket_distribution;
700
    // TODO: Create CacheOperator is confused now
701
663k
    OperatorPtr op = nullptr;
702
663k
    OperatorPtr cache_op = nullptr;
703
663k
    RETURN_IF_ERROR(_create_operator(pool, tnodes[*node_idx], descs, op, cur_pipe,
704
663k
                                     parent == nullptr ? -1 : parent->node_id(), child_idx,
705
663k
                                     followed_by_shuffled_operator,
706
663k
                                     current_require_bucket_distribution, cache_op));
707
    // Initialization must be done here. For example, group by expressions in agg will be used to
708
    // decide if a local shuffle should be planed, so it must be initialized here.
709
663k
    RETURN_IF_ERROR(op->init(tnode, _runtime_state.get()));
710
    // assert(parent != nullptr || (node_idx == 0 && root_expr != nullptr));
711
663k
    if (parent != nullptr) {
712
        // add to parent's child(s)
713
236k
        RETURN_IF_ERROR(parent->set_child(cache_op ? cache_op : op));
714
426k
    } else {
715
426k
        *root = op;
716
426k
    }
717
    /**
718
     * `ExchangeType::HASH_SHUFFLE` should be used if an operator is followed by a shuffled operator (shuffled hash join, union operator followed by co-located operators).
719
     *
720
     * For plan:
721
     * LocalExchange(id=0) -> Aggregation(id=1) -> ShuffledHashJoin(id=2)
722
     *                           Exchange(id=3) -> ShuffledHashJoinBuild(id=2)
723
     * We must ensure data distribution of `LocalExchange(id=0)` is same as Exchange(id=3).
724
     *
725
     * If an operator's is followed by a local exchange without shuffle (e.g. passthrough), a
726
     * shuffled local exchanger will be used before join so it is not followed by shuffle join.
727
     */
728
663k
    auto required_data_distribution =
729
663k
            cur_pipe->operators().empty()
730
663k
                    ? cur_pipe->sink()->required_data_distribution(_runtime_state.get())
731
663k
                    : op->required_data_distribution(_runtime_state.get());
732
663k
    current_followed_by_shuffled_operator =
733
663k
            ((followed_by_shuffled_operator ||
734
663k
              (cur_pipe->operators().empty() ? cur_pipe->sink()->is_shuffled_operator()
735
607k
                                             : op->is_shuffled_operator())) &&
736
663k
             Pipeline::is_hash_exchange(required_data_distribution.distribution_type)) ||
737
663k
            (followed_by_shuffled_operator &&
738
554k
             required_data_distribution.distribution_type == ExchangeType::NOOP);
739
740
663k
    current_require_bucket_distribution =
741
663k
            ((require_bucket_distribution ||
742
663k
              (cur_pipe->operators().empty() ? cur_pipe->sink()->is_colocated_operator()
743
611k
                                             : op->is_colocated_operator())) &&
744
663k
             Pipeline::is_hash_exchange(required_data_distribution.distribution_type)) ||
745
663k
            (require_bucket_distribution &&
746
561k
             required_data_distribution.distribution_type == ExchangeType::NOOP);
747
748
663k
    if (num_children == 0) {
749
445k
        _use_serial_source = op->is_serial_operator();
750
445k
    }
751
    // rely on that tnodes is preorder of the plan
752
901k
    for (int i = 0; i < num_children; i++) {
753
237k
        ++*node_idx;
754
237k
        RETURN_IF_ERROR(_create_tree_helper(pool, tnodes, descs, op, node_idx, nullptr, cur_pipe, i,
755
237k
                                            current_followed_by_shuffled_operator,
756
237k
                                            current_require_bucket_distribution));
757
758
        // we are expecting a child, but have used all nodes
759
        // this means we have been given a bad tree and must fail
760
237k
        if (*node_idx >= tnodes.size()) {
761
0
            return Status::InternalError(
762
0
                    "Failed to reconstruct plan tree from thrift. Node id: {}, number of "
763
0
                    "nodes: {}",
764
0
                    *node_idx, tnodes.size());
765
0
        }
766
237k
    }
767
768
663k
    return Status::OK();
769
663k
}
770
771
void PipelineFragmentContext::_inherit_pipeline_properties(
772
        const DataDistribution& data_distribution, PipelinePtr pipe_with_source,
773
111k
        PipelinePtr pipe_with_sink) {
774
111k
    pipe_with_sink->set_num_tasks(pipe_with_source->num_tasks());
775
111k
    pipe_with_source->set_num_tasks(_num_instances);
776
111k
    pipe_with_source->set_data_distribution(data_distribution);
777
111k
}
778
779
Status PipelineFragmentContext::_add_local_exchange_impl(
780
        int idx, ObjectPool* pool, PipelinePtr cur_pipe, PipelinePtr new_pip,
781
        DataDistribution data_distribution, bool* do_local_exchange, int num_buckets,
782
        const std::map<int, int>& bucket_seq_to_instance_idx,
783
111k
        const std::map<int, int>& shuffle_idx_to_instance_idx) {
784
111k
    auto& operators = cur_pipe->operators();
785
111k
    const auto downstream_pipeline_id = cur_pipe->id();
786
111k
    auto local_exchange_id = next_operator_id();
787
    // 1. Create a new pipeline with local exchange sink.
788
111k
    DataSinkOperatorPtr sink;
789
111k
    auto sink_id = next_sink_operator_id();
790
791
    /**
792
     * `bucket_seq_to_instance_idx` is empty if no scan operator is contained in this fragment.
793
     * So co-located operators(e.g. Agg, Analytic) should use `HASH_SHUFFLE` instead of `BUCKET_HASH_SHUFFLE`.
794
     */
795
111k
    const bool followed_by_shuffled_operator =
796
111k
            operators.size() > idx ? operators[idx]->followed_by_shuffled_operator()
797
111k
                                   : cur_pipe->sink()->followed_by_shuffled_operator();
798
111k
    const bool use_global_hash_shuffle = bucket_seq_to_instance_idx.empty() &&
799
111k
                                         !shuffle_idx_to_instance_idx.contains(-1) &&
800
111k
                                         followed_by_shuffled_operator && !_use_serial_source;
801
111k
    sink = std::make_shared<LocalExchangeSinkOperatorX>(
802
111k
            sink_id, local_exchange_id, use_global_hash_shuffle ? _total_instances : _num_instances,
803
111k
            data_distribution.partition_exprs, bucket_seq_to_instance_idx);
804
111k
    if (bucket_seq_to_instance_idx.empty() &&
805
111k
        data_distribution.distribution_type == ExchangeType::BUCKET_HASH_SHUFFLE) {
806
4
        data_distribution.distribution_type = ExchangeType::HASH_SHUFFLE;
807
4
    }
808
111k
    RETURN_IF_ERROR(new_pip->set_sink(sink));
809
111k
    RETURN_IF_ERROR(new_pip->sink()->init(_runtime_state.get(), data_distribution.distribution_type,
810
111k
                                          num_buckets, use_global_hash_shuffle,
811
111k
                                          shuffle_idx_to_instance_idx));
812
813
    // 2. Create and initialize LocalExchangeSharedState.
814
111k
    std::shared_ptr<LocalExchangeSharedState> shared_state =
815
111k
            LocalExchangeSharedState::create_shared(_num_instances);
816
111k
    switch (data_distribution.distribution_type) {
817
18.9k
    case ExchangeType::HASH_SHUFFLE:
818
18.9k
        shared_state->exchanger = ShuffleExchanger::create_unique(
819
18.9k
                std::max(cur_pipe->num_tasks(), _num_instances), _num_instances,
820
18.9k
                use_global_hash_shuffle ? _total_instances : _num_instances,
821
18.9k
                _runtime_state->query_options().__isset.local_exchange_free_blocks_limit
822
18.9k
                        ? cast_set<int>(
823
18.9k
                                  _runtime_state->query_options().local_exchange_free_blocks_limit)
824
18.9k
                        : 0);
825
18.9k
        break;
826
501
    case ExchangeType::BUCKET_HASH_SHUFFLE:
827
501
        shared_state->exchanger = BucketShuffleExchanger::create_unique(
828
501
                std::max(cur_pipe->num_tasks(), _num_instances), _num_instances, num_buckets,
829
501
                _runtime_state->query_options().__isset.local_exchange_free_blocks_limit
830
501
                        ? cast_set<int>(
831
501
                                  _runtime_state->query_options().local_exchange_free_blocks_limit)
832
501
                        : 0);
833
501
        break;
834
87.9k
    case ExchangeType::PASSTHROUGH:
835
87.9k
        shared_state->exchanger = PassthroughExchanger::create_unique(
836
87.9k
                cur_pipe->num_tasks(), _num_instances,
837
87.9k
                _runtime_state->query_options().__isset.local_exchange_free_blocks_limit
838
87.9k
                        ? cast_set<int>(
839
87.9k
                                  _runtime_state->query_options().local_exchange_free_blocks_limit)
840
87.9k
                        : 0);
841
87.9k
        break;
842
315
    case ExchangeType::BROADCAST:
843
315
        shared_state->exchanger = BroadcastExchanger::create_unique(
844
315
                cur_pipe->num_tasks(), _num_instances,
845
315
                _runtime_state->query_options().__isset.local_exchange_free_blocks_limit
846
315
                        ? cast_set<int>(
847
315
                                  _runtime_state->query_options().local_exchange_free_blocks_limit)
848
315
                        : 0);
849
315
        break;
850
2.75k
    case ExchangeType::PASS_TO_ONE:
851
2.75k
        if (_runtime_state->enable_share_hash_table_for_broadcast_join()) {
852
            // If shared hash table is enabled for BJ, hash table will be built by only one task
853
1.74k
            shared_state->exchanger = PassToOneExchanger::create_unique(
854
1.74k
                    cur_pipe->num_tasks(), _num_instances,
855
1.74k
                    _runtime_state->query_options().__isset.local_exchange_free_blocks_limit
856
1.74k
                            ? cast_set<int>(_runtime_state->query_options()
857
1.74k
                                                    .local_exchange_free_blocks_limit)
858
1.74k
                            : 0);
859
1.74k
        } else {
860
1.00k
            shared_state->exchanger = BroadcastExchanger::create_unique(
861
1.00k
                    cur_pipe->num_tasks(), _num_instances,
862
1.00k
                    _runtime_state->query_options().__isset.local_exchange_free_blocks_limit
863
1.00k
                            ? cast_set<int>(_runtime_state->query_options()
864
1.00k
                                                    .local_exchange_free_blocks_limit)
865
1.00k
                            : 0);
866
1.00k
        }
867
2.75k
        break;
868
916
    case ExchangeType::ADAPTIVE_PASSTHROUGH:
869
916
        shared_state->exchanger = AdaptivePassthroughExchanger::create_unique(
870
916
                std::max(cur_pipe->num_tasks(), _num_instances), _num_instances,
871
916
                _runtime_state->query_options().__isset.local_exchange_free_blocks_limit
872
916
                        ? cast_set<int>(
873
916
                                  _runtime_state->query_options().local_exchange_free_blocks_limit)
874
916
                        : 0);
875
916
        break;
876
0
    default:
877
0
        return Status::InternalError("Unsupported local exchange type : " +
878
0
                                     std::to_string((int)data_distribution.distribution_type));
879
111k
    }
880
111k
    shared_state->create_source_dependencies(_num_instances, local_exchange_id, local_exchange_id,
881
111k
                                             "LOCAL_EXCHANGE_OPERATOR");
882
111k
    shared_state->create_sink_dependency(sink_id, local_exchange_id, "LOCAL_EXCHANGE_SINK");
883
111k
    _op_id_to_shared_state.insert({local_exchange_id, {shared_state, shared_state->sink_deps}});
884
885
    // 3. Set two pipelines' operator list. For example, split pipeline [Scan - AggSink] to
886
    // pipeline1 [Scan - LocalExchangeSink] and pipeline2 [LocalExchangeSource - AggSink].
887
888
    // 3.1 Initialize new pipeline's operator list.
889
111k
    std::copy(operators.begin(), operators.begin() + idx,
890
111k
              std::inserter(new_pip->operators(), new_pip->operators().end()));
891
892
    // 3.2 Erase unused operators in previous pipeline.
893
111k
    operators.erase(operators.begin(), operators.begin() + idx);
894
895
    // 4. Initialize LocalExchangeSource and insert it into this pipeline.
896
111k
    OperatorPtr source_op;
897
111k
    source_op = std::make_shared<LocalExchangeSourceOperatorX>(pool, local_exchange_id);
898
111k
    RETURN_IF_ERROR(source_op->set_child(new_pip->operators().back()));
899
111k
    RETURN_IF_ERROR(source_op->init(data_distribution.distribution_type));
900
111k
    if (!operators.empty()) {
901
43.8k
        RETURN_IF_ERROR(operators.front()->set_child(nullptr));
902
43.8k
        RETURN_IF_ERROR(operators.front()->set_child(source_op));
903
43.8k
    }
904
111k
    operators.insert(operators.begin(), source_op);
905
906
    // 5. Set children for two pipelines separately.
907
111k
    std::vector<std::shared_ptr<Pipeline>> new_children;
908
111k
    std::vector<PipelineId> edges_with_source;
909
130k
    for (auto child : cur_pipe->children()) {
910
130k
        bool found = false;
911
144k
        for (auto op : new_pip->operators()) {
912
144k
            if (child->sink()->node_id() == op->node_id()) {
913
13.0k
                new_pip->set_children(child);
914
13.0k
                found = true;
915
13.0k
            };
916
144k
        }
917
130k
        if (!found) {
918
117k
            new_children.push_back(child);
919
117k
            edges_with_source.push_back(child->id());
920
117k
        }
921
130k
    }
922
111k
    new_children.push_back(new_pip);
923
111k
    edges_with_source.push_back(new_pip->id());
924
925
    // 6. Set DAG for new pipelines.
926
111k
    if (!new_pip->children().empty()) {
927
7.21k
        std::vector<PipelineId> edges_with_sink;
928
13.0k
        for (auto child : new_pip->children()) {
929
13.0k
            edges_with_sink.push_back(child->id());
930
13.0k
        }
931
7.21k
        _dag.insert({new_pip->id(), edges_with_sink});
932
7.21k
    }
933
111k
    cur_pipe->set_children(new_children);
934
111k
    _dag[downstream_pipeline_id] = edges_with_source;
935
111k
    RETURN_IF_ERROR(new_pip->sink()->set_child(new_pip->operators().back()));
936
111k
    RETURN_IF_ERROR(cur_pipe->sink()->set_child(nullptr));
937
111k
    RETURN_IF_ERROR(cur_pipe->sink()->set_child(cur_pipe->operators().back()));
938
939
    // 7. Inherit properties from current pipeline.
940
111k
    _inherit_pipeline_properties(data_distribution, cur_pipe, new_pip);
941
111k
    return Status::OK();
942
111k
}
943
944
Status PipelineFragmentContext::_add_local_exchange(
945
        int pip_idx, int idx, int node_id, ObjectPool* pool, PipelinePtr cur_pipe,
946
        DataDistribution data_distribution, bool* do_local_exchange, int num_buckets,
947
        const std::map<int, int>& bucket_seq_to_instance_idx,
948
191k
        const std::map<int, int>& shuffle_idx_to_instance_idx) {
949
191k
    if (_num_instances <= 1 || cur_pipe->num_tasks_of_parent() <= 1) {
950
52.6k
        return Status::OK();
951
52.6k
    }
952
953
139k
    if (!cur_pipe->need_to_local_exchange(data_distribution, idx)) {
954
47.1k
        return Status::OK();
955
47.1k
    }
956
92.1k
    *do_local_exchange = true;
957
958
92.1k
    auto& operators = cur_pipe->operators();
959
92.1k
    auto total_op_num = operators.size();
960
92.1k
    auto new_pip = add_pipeline(cur_pipe, pip_idx + 1);
961
92.1k
    RETURN_IF_ERROR(_add_local_exchange_impl(
962
92.1k
            idx, pool, cur_pipe, new_pip, data_distribution, do_local_exchange, num_buckets,
963
92.1k
            bucket_seq_to_instance_idx, shuffle_idx_to_instance_idx));
964
965
18.4E
    CHECK(total_op_num + 1 == cur_pipe->operators().size() + new_pip->operators().size())
966
18.4E
            << "total_op_num: " << total_op_num
967
18.4E
            << " cur_pipe->operators().size(): " << cur_pipe->operators().size()
968
18.4E
            << " new_pip->operators().size(): " << new_pip->operators().size();
969
970
    // There are some local shuffles with relatively heavy operations on the sink.
971
    // If the local sink concurrency is 1 and the local source concurrency is n, the sink becomes a bottleneck.
972
    // Therefore, local passthrough is used to increase the concurrency of the sink.
973
    // op -> local sink(1) -> local source (n)
974
    // op -> local passthrough(1) -> local passthrough(n) ->  local sink(n) -> local source (n)
975
92.2k
    if (cur_pipe->num_tasks() > 1 && new_pip->num_tasks() == 1 &&
976
92.1k
        Pipeline::heavy_operations_on_the_sink(data_distribution.distribution_type)) {
977
19.0k
        RETURN_IF_ERROR(_add_local_exchange_impl(
978
19.0k
                cast_set<int>(new_pip->operators().size()), pool, new_pip,
979
19.0k
                add_pipeline(new_pip, pip_idx + 2), DataDistribution(ExchangeType::PASSTHROUGH),
980
19.0k
                do_local_exchange, num_buckets, bucket_seq_to_instance_idx,
981
19.0k
                shuffle_idx_to_instance_idx));
982
19.0k
    }
983
92.1k
    return Status::OK();
984
92.1k
}
985
986
Status PipelineFragmentContext::_plan_local_exchange(
987
        int num_buckets, const std::map<int, int>& bucket_seq_to_instance_idx,
988
426k
        const std::map<int, int>& shuffle_idx_to_instance_idx) {
989
983k
    for (int pip_idx = cast_set<int>(_pipelines.size()) - 1; pip_idx >= 0; pip_idx--) {
990
557k
        _pipelines[pip_idx]->init_data_distribution(_runtime_state.get());
991
        // Set property if child pipeline is not join operator's child.
992
557k
        if (!_pipelines[pip_idx]->children().empty()) {
993
124k
            for (auto& child : _pipelines[pip_idx]->children()) {
994
124k
                if (child->sink()->node_id() ==
995
124k
                    _pipelines[pip_idx]->operators().front()->node_id()) {
996
109k
                    _pipelines[pip_idx]->set_data_distribution(child->data_distribution());
997
109k
                }
998
124k
            }
999
119k
        }
1000
1001
        // if 'num_buckets == 0' means the fragment is colocated by exchange node not the
1002
        // scan node. so here use `_num_instance` to replace the `num_buckets` to prevent dividing 0
1003
        // still keep colocate plan after local shuffle
1004
557k
        RETURN_IF_ERROR(_plan_local_exchange(num_buckets, pip_idx, _pipelines[pip_idx],
1005
557k
                                             bucket_seq_to_instance_idx,
1006
557k
                                             shuffle_idx_to_instance_idx));
1007
557k
    }
1008
426k
    return Status::OK();
1009
426k
}
1010
1011
Status PipelineFragmentContext::_plan_local_exchange(
1012
        int num_buckets, int pip_idx, PipelinePtr pip,
1013
        const std::map<int, int>& bucket_seq_to_instance_idx,
1014
556k
        const std::map<int, int>& shuffle_idx_to_instance_idx) {
1015
556k
    int idx = 1;
1016
556k
    bool do_local_exchange = false;
1017
600k
    do {
1018
600k
        auto& ops = pip->operators();
1019
600k
        do_local_exchange = false;
1020
        // Plan local exchange for each operator.
1021
669k
        for (; idx < ops.size();) {
1022
113k
            if (ops[idx]->required_data_distribution(_runtime_state.get()).need_local_exchange()) {
1023
105k
                RETURN_IF_ERROR(_add_local_exchange(
1024
105k
                        pip_idx, idx, ops[idx]->node_id(), _runtime_state->obj_pool(), pip,
1025
105k
                        ops[idx]->required_data_distribution(_runtime_state.get()),
1026
105k
                        &do_local_exchange, num_buckets, bucket_seq_to_instance_idx,
1027
105k
                        shuffle_idx_to_instance_idx));
1028
105k
            }
1029
113k
            if (do_local_exchange) {
1030
                // If local exchange is needed for current operator, we will split this pipeline to
1031
                // two pipelines by local exchange sink/source. And then we need to process remaining
1032
                // operators in this pipeline so we set idx to 2 (0 is local exchange source and 1
1033
                // is current operator was already processed) and continue to plan local exchange.
1034
43.9k
                idx = 2;
1035
43.9k
                break;
1036
43.9k
            }
1037
69.2k
            idx++;
1038
69.2k
        }
1039
600k
    } while (do_local_exchange);
1040
556k
    if (pip->sink()->required_data_distribution(_runtime_state.get()).need_local_exchange()) {
1041
86.1k
        RETURN_IF_ERROR(_add_local_exchange(
1042
86.1k
                pip_idx, idx, pip->sink()->node_id(), _runtime_state->obj_pool(), pip,
1043
86.1k
                pip->sink()->required_data_distribution(_runtime_state.get()), &do_local_exchange,
1044
86.1k
                num_buckets, bucket_seq_to_instance_idx, shuffle_idx_to_instance_idx));
1045
86.1k
    }
1046
556k
    return Status::OK();
1047
556k
}
1048
1049
Status PipelineFragmentContext::_create_data_sink(ObjectPool* pool, const TDataSink& thrift_sink,
1050
                                                  const std::vector<TExpr>& output_exprs,
1051
                                                  const TPipelineFragmentParams& params,
1052
                                                  const RowDescriptor& row_desc,
1053
                                                  RuntimeState* state, DescriptorTbl& desc_tbl,
1054
429k
                                                  PipelineId cur_pipeline_id) {
1055
429k
    switch (thrift_sink.type) {
1056
144k
    case TDataSinkType::DATA_STREAM_SINK: {
1057
144k
        if (!thrift_sink.__isset.stream_sink) {
1058
0
            return Status::InternalError("Missing data stream sink.");
1059
0
        }
1060
144k
        _sink = std::make_shared<ExchangeSinkOperatorX>(
1061
144k
                state, row_desc, next_sink_operator_id(), thrift_sink.stream_sink,
1062
144k
                params.destinations, _fragment_instance_ids);
1063
144k
        break;
1064
144k
    }
1065
249k
    case TDataSinkType::RESULT_SINK: {
1066
249k
        if (!thrift_sink.__isset.result_sink) {
1067
0
            return Status::InternalError("Missing data buffer sink.");
1068
0
        }
1069
1070
249k
        _sink = std::make_shared<ResultSinkOperatorX>(next_sink_operator_id(), row_desc,
1071
249k
                                                      output_exprs, thrift_sink.result_sink);
1072
249k
        break;
1073
249k
    }
1074
105
    case TDataSinkType::DICTIONARY_SINK: {
1075
105
        if (!thrift_sink.__isset.dictionary_sink) {
1076
0
            return Status::InternalError("Missing dict sink.");
1077
0
        }
1078
1079
105
        _sink = std::make_shared<DictSinkOperatorX>(next_sink_operator_id(), row_desc, output_exprs,
1080
105
                                                    thrift_sink.dictionary_sink);
1081
105
        break;
1082
105
    }
1083
0
    case TDataSinkType::GROUP_COMMIT_OLAP_TABLE_SINK:
1084
30.1k
    case TDataSinkType::OLAP_TABLE_SINK: {
1085
30.1k
        if (state->query_options().enable_memtable_on_sink_node &&
1086
30.1k
            !_has_inverted_index_v1_or_partial_update(thrift_sink.olap_table_sink) &&
1087
30.1k
            !config::is_cloud_mode()) {
1088
2.10k
            _sink = std::make_shared<OlapTableSinkV2OperatorX>(pool, next_sink_operator_id(),
1089
2.10k
                                                               row_desc, output_exprs);
1090
28.0k
        } else {
1091
28.0k
            _sink = std::make_shared<OlapTableSinkOperatorX>(pool, next_sink_operator_id(),
1092
28.0k
                                                             row_desc, output_exprs);
1093
28.0k
        }
1094
30.1k
        break;
1095
0
    }
1096
165
    case TDataSinkType::GROUP_COMMIT_BLOCK_SINK: {
1097
165
        DCHECK(thrift_sink.__isset.olap_table_sink);
1098
165
        DCHECK(state->get_query_ctx() != nullptr);
1099
165
        state->get_query_ctx()->query_mem_tracker()->is_group_commit_load = true;
1100
165
        _sink = std::make_shared<GroupCommitBlockSinkOperatorX>(next_sink_operator_id(), row_desc,
1101
165
                                                                output_exprs);
1102
165
        break;
1103
0
    }
1104
1.46k
    case TDataSinkType::HIVE_TABLE_SINK: {
1105
1.46k
        if (!thrift_sink.__isset.hive_table_sink) {
1106
0
            return Status::InternalError("Missing hive table sink.");
1107
0
        }
1108
1.46k
        _sink = std::make_shared<HiveTableSinkOperatorX>(pool, next_sink_operator_id(), row_desc,
1109
1.46k
                                                         output_exprs);
1110
1.46k
        break;
1111
1.46k
    }
1112
1.73k
    case TDataSinkType::ICEBERG_TABLE_SINK: {
1113
1.73k
        if (!thrift_sink.__isset.iceberg_table_sink) {
1114
0
            return Status::InternalError("Missing iceberg table sink.");
1115
0
        }
1116
1.73k
        if (thrift_sink.iceberg_table_sink.__isset.sort_info) {
1117
0
            _sink = std::make_shared<SpillIcebergTableSinkOperatorX>(pool, next_sink_operator_id(),
1118
0
                                                                     row_desc, output_exprs);
1119
1.73k
        } else {
1120
1.73k
            _sink = std::make_shared<IcebergTableSinkOperatorX>(pool, next_sink_operator_id(),
1121
1.73k
                                                                row_desc, output_exprs);
1122
1.73k
        }
1123
1.73k
        break;
1124
1.73k
    }
1125
20
    case TDataSinkType::ICEBERG_DELETE_SINK: {
1126
20
        if (!thrift_sink.__isset.iceberg_delete_sink) {
1127
0
            return Status::InternalError("Missing iceberg delete sink.");
1128
0
        }
1129
20
        _sink = std::make_shared<IcebergDeleteSinkOperatorX>(pool, next_sink_operator_id(),
1130
20
                                                             row_desc, output_exprs);
1131
20
        break;
1132
20
    }
1133
80
    case TDataSinkType::ICEBERG_MERGE_SINK: {
1134
80
        if (!thrift_sink.__isset.iceberg_merge_sink) {
1135
0
            return Status::InternalError("Missing iceberg merge sink.");
1136
0
        }
1137
80
        _sink = std::make_shared<IcebergMergeSinkOperatorX>(pool, next_sink_operator_id(), row_desc,
1138
80
                                                            output_exprs);
1139
80
        break;
1140
80
    }
1141
0
    case TDataSinkType::MAXCOMPUTE_TABLE_SINK: {
1142
0
        if (!thrift_sink.__isset.max_compute_table_sink) {
1143
0
            return Status::InternalError("Missing max compute table sink.");
1144
0
        }
1145
0
        _sink = std::make_shared<MCTableSinkOperatorX>(pool, next_sink_operator_id(), row_desc,
1146
0
                                                       output_exprs);
1147
0
        break;
1148
0
    }
1149
80
    case TDataSinkType::JDBC_TABLE_SINK: {
1150
80
        if (!thrift_sink.__isset.jdbc_table_sink) {
1151
0
            return Status::InternalError("Missing data jdbc sink.");
1152
0
        }
1153
80
        if (config::enable_java_support) {
1154
80
            _sink = std::make_shared<JdbcTableSinkOperatorX>(row_desc, next_sink_operator_id(),
1155
80
                                                             output_exprs);
1156
80
        } else {
1157
0
            return Status::InternalError(
1158
0
                    "Jdbc table sink is not enabled, you can change be config "
1159
0
                    "enable_java_support to true and restart be.");
1160
0
        }
1161
80
        break;
1162
80
    }
1163
80
    case TDataSinkType::MEMORY_SCRATCH_SINK: {
1164
3
        if (!thrift_sink.__isset.memory_scratch_sink) {
1165
0
            return Status::InternalError("Missing data buffer sink.");
1166
0
        }
1167
1168
3
        _sink = std::make_shared<MemoryScratchSinkOperatorX>(row_desc, next_sink_operator_id(),
1169
3
                                                             output_exprs);
1170
3
        break;
1171
3
    }
1172
502
    case TDataSinkType::RESULT_FILE_SINK: {
1173
502
        if (!thrift_sink.__isset.result_file_sink) {
1174
0
            return Status::InternalError("Missing result file sink.");
1175
0
        }
1176
1177
        // Result file sink is not the top sink
1178
502
        if (params.__isset.destinations && !params.destinations.empty()) {
1179
0
            _sink = std::make_shared<ResultFileSinkOperatorX>(
1180
0
                    next_sink_operator_id(), row_desc, thrift_sink.result_file_sink,
1181
0
                    params.destinations, output_exprs, desc_tbl);
1182
502
        } else {
1183
502
            _sink = std::make_shared<ResultFileSinkOperatorX>(next_sink_operator_id(), row_desc,
1184
502
                                                              output_exprs);
1185
502
        }
1186
502
        break;
1187
502
    }
1188
1.95k
    case TDataSinkType::MULTI_CAST_DATA_STREAM_SINK: {
1189
1.95k
        DCHECK(thrift_sink.__isset.multi_cast_stream_sink);
1190
1.95k
        DCHECK_GT(thrift_sink.multi_cast_stream_sink.sinks.size(), 0);
1191
1.95k
        auto sink_id = next_sink_operator_id();
1192
1.95k
        const int multi_cast_node_id = sink_id;
1193
1.95k
        auto sender_size = thrift_sink.multi_cast_stream_sink.sinks.size();
1194
        // one sink has multiple sources.
1195
1.95k
        std::vector<int> sources;
1196
7.59k
        for (int i = 0; i < sender_size; ++i) {
1197
5.64k
            auto source_id = next_operator_id();
1198
5.64k
            sources.push_back(source_id);
1199
5.64k
        }
1200
1201
1.95k
        _sink = std::make_shared<MultiCastDataStreamSinkOperatorX>(
1202
1.95k
                sink_id, multi_cast_node_id, sources, pool, thrift_sink.multi_cast_stream_sink);
1203
7.59k
        for (int i = 0; i < sender_size; ++i) {
1204
5.64k
            auto new_pipeline = add_pipeline();
1205
            // use to exchange sink
1206
5.64k
            RowDescriptor* exchange_row_desc = nullptr;
1207
5.64k
            {
1208
5.64k
                const auto& tmp_row_desc =
1209
5.64k
                        !thrift_sink.multi_cast_stream_sink.sinks[i].output_exprs.empty()
1210
5.64k
                                ? RowDescriptor(state->desc_tbl(),
1211
5.64k
                                                {thrift_sink.multi_cast_stream_sink.sinks[i]
1212
5.64k
                                                         .output_tuple_id})
1213
5.64k
                                : row_desc;
1214
5.64k
                exchange_row_desc = pool->add(new RowDescriptor(tmp_row_desc));
1215
5.64k
            }
1216
5.64k
            auto source_id = sources[i];
1217
5.64k
            OperatorPtr source_op;
1218
            // 1. create and set the source operator of multi_cast_data_stream_source for new pipeline
1219
5.64k
            source_op = std::make_shared<MultiCastDataStreamerSourceOperatorX>(
1220
5.64k
                    /*node_id*/ source_id, /*consumer_id*/ i, pool,
1221
5.64k
                    thrift_sink.multi_cast_stream_sink.sinks[i], row_desc,
1222
5.64k
                    /*operator_id=*/source_id);
1223
5.64k
            RETURN_IF_ERROR(new_pipeline->add_operator(
1224
5.64k
                    source_op, params.__isset.parallel_instances ? params.parallel_instances : 0));
1225
            // 2. create and set sink operator of data stream sender for new pipeline
1226
1227
5.64k
            DataSinkOperatorPtr sink_op;
1228
5.64k
            sink_op = std::make_shared<ExchangeSinkOperatorX>(
1229
5.64k
                    state, *exchange_row_desc, next_sink_operator_id(),
1230
5.64k
                    thrift_sink.multi_cast_stream_sink.sinks[i],
1231
5.64k
                    thrift_sink.multi_cast_stream_sink.destinations[i], _fragment_instance_ids);
1232
1233
5.64k
            RETURN_IF_ERROR(new_pipeline->set_sink(sink_op));
1234
5.64k
            {
1235
5.64k
                TDataSink* t = pool->add(new TDataSink());
1236
5.64k
                t->stream_sink = thrift_sink.multi_cast_stream_sink.sinks[i];
1237
5.64k
                RETURN_IF_ERROR(sink_op->init(*t));
1238
5.64k
            }
1239
1240
            // 3. set dependency dag
1241
5.64k
            _dag[new_pipeline->id()].push_back(cur_pipeline_id);
1242
5.64k
        }
1243
1.95k
        if (sources.empty()) {
1244
0
            return Status::InternalError("size of sources must be greater than 0");
1245
0
        }
1246
1.95k
        break;
1247
1.95k
    }
1248
1.95k
    case TDataSinkType::BLACKHOLE_SINK: {
1249
13
        if (!thrift_sink.__isset.blackhole_sink) {
1250
0
            return Status::InternalError("Missing blackhole sink.");
1251
0
        }
1252
1253
13
        _sink.reset(new BlackholeSinkOperatorX(next_sink_operator_id()));
1254
13
        break;
1255
13
    }
1256
156
    case TDataSinkType::TVF_TABLE_SINK: {
1257
156
        if (!thrift_sink.__isset.tvf_table_sink) {
1258
0
            return Status::InternalError("Missing TVF table sink.");
1259
0
        }
1260
156
        _sink = std::make_shared<TVFTableSinkOperatorX>(pool, next_sink_operator_id(), row_desc,
1261
156
                                                        output_exprs);
1262
156
        break;
1263
156
    }
1264
0
    default:
1265
0
        return Status::InternalError("Unsuported sink type in pipeline: {}", thrift_sink.type);
1266
429k
    }
1267
428k
    return Status::OK();
1268
429k
}
1269
1270
// NOLINTBEGIN(readability-function-size)
1271
// NOLINTBEGIN(readability-function-cognitive-complexity)
1272
Status PipelineFragmentContext::_create_operator(ObjectPool* pool, const TPlanNode& tnode,
1273
                                                 const DescriptorTbl& descs, OperatorPtr& op,
1274
                                                 PipelinePtr& cur_pipe, int parent_idx,
1275
                                                 int child_idx,
1276
                                                 const bool followed_by_shuffled_operator,
1277
                                                 const bool require_bucket_distribution,
1278
667k
                                                 OperatorPtr& cache_op) {
1279
667k
    std::vector<DataSinkOperatorPtr> sink_ops;
1280
667k
    Defer defer = Defer([&]() {
1281
666k
        if (op) {
1282
666k
            op->update_operator(tnode, followed_by_shuffled_operator, require_bucket_distribution);
1283
666k
        }
1284
666k
        for (auto& s : sink_ops) {
1285
124k
            s->update_operator(tnode, followed_by_shuffled_operator, require_bucket_distribution);
1286
124k
        }
1287
666k
    });
1288
    // We directly construct the operator from Thrift because the given array is in the order of preorder traversal.
1289
    // Therefore, here we need to use a stack-like structure.
1290
667k
    _pipeline_parent_map.pop(cur_pipe, parent_idx, child_idx);
1291
667k
    std::stringstream error_msg;
1292
667k
    bool enable_query_cache = _params.fragment.__isset.query_cache_param;
1293
1294
667k
    bool fe_with_old_version = false;
1295
667k
    switch (tnode.node_type) {
1296
210k
    case TPlanNodeType::OLAP_SCAN_NODE: {
1297
210k
        op = std::make_shared<OlapScanOperatorX>(
1298
210k
                pool, tnode, next_operator_id(), descs, _num_instances,
1299
210k
                enable_query_cache ? _params.fragment.query_cache_param : TQueryCacheParam {});
1300
210k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1301
210k
        fe_with_old_version = !tnode.__isset.is_serial_operator;
1302
210k
        break;
1303
210k
    }
1304
78
    case TPlanNodeType::GROUP_COMMIT_SCAN_NODE: {
1305
78
        DCHECK(_query_ctx != nullptr);
1306
78
        _query_ctx->query_mem_tracker()->is_group_commit_load = true;
1307
78
        op = std::make_shared<GroupCommitOperatorX>(pool, tnode, next_operator_id(), descs,
1308
78
                                                    _num_instances);
1309
78
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1310
78
        fe_with_old_version = !tnode.__isset.is_serial_operator;
1311
78
        break;
1312
78
    }
1313
0
    case TPlanNodeType::JDBC_SCAN_NODE: {
1314
0
        if (config::enable_java_support) {
1315
0
            op = std::make_shared<JDBCScanOperatorX>(pool, tnode, next_operator_id(), descs,
1316
0
                                                     _num_instances);
1317
0
            RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1318
0
        } else {
1319
0
            return Status::InternalError(
1320
0
                    "Jdbc scan node is disabled, you can change be config enable_java_support "
1321
0
                    "to true and restart be.");
1322
0
        }
1323
0
        fe_with_old_version = !tnode.__isset.is_serial_operator;
1324
0
        break;
1325
0
    }
1326
23.0k
    case TPlanNodeType::FILE_SCAN_NODE: {
1327
23.0k
        op = std::make_shared<FileScanOperatorX>(pool, tnode, next_operator_id(), descs,
1328
23.0k
                                                 _num_instances);
1329
23.0k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1330
23.0k
        fe_with_old_version = !tnode.__isset.is_serial_operator;
1331
23.0k
        break;
1332
23.0k
    }
1333
0
    case TPlanNodeType::ES_SCAN_NODE:
1334
592
    case TPlanNodeType::ES_HTTP_SCAN_NODE: {
1335
592
        op = std::make_shared<EsScanOperatorX>(pool, tnode, next_operator_id(), descs,
1336
592
                                               _num_instances);
1337
592
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1338
592
        fe_with_old_version = !tnode.__isset.is_serial_operator;
1339
592
        break;
1340
592
    }
1341
147k
    case TPlanNodeType::EXCHANGE_NODE: {
1342
147k
        int num_senders = _params.per_exch_num_senders.contains(tnode.node_id)
1343
147k
                                  ? _params.per_exch_num_senders.find(tnode.node_id)->second
1344
18.4E
                                  : 0;
1345
147k
        DCHECK_GT(num_senders, 0);
1346
147k
        op = std::make_shared<ExchangeSourceOperatorX>(pool, tnode, next_operator_id(), descs,
1347
147k
                                                       num_senders);
1348
147k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1349
147k
        fe_with_old_version = !tnode.__isset.is_serial_operator;
1350
147k
        break;
1351
147k
    }
1352
156k
    case TPlanNodeType::AGGREGATION_NODE: {
1353
156k
        if (tnode.agg_node.grouping_exprs.empty() &&
1354
156k
            descs.get_tuple_descriptor(tnode.agg_node.output_tuple_id)->slots().empty()) {
1355
0
            return Status::InternalError("Illegal aggregate node " + std::to_string(tnode.node_id) +
1356
0
                                         ": group by and output is empty");
1357
0
        }
1358
156k
        bool need_create_cache_op =
1359
156k
                enable_query_cache && tnode.node_id == _params.fragment.query_cache_param.node_id;
1360
156k
        auto create_query_cache_operator = [&](PipelinePtr& new_pipe) {
1361
10
            auto cache_node_id = _params.local_params[0].per_node_scan_ranges.begin()->first;
1362
10
            auto cache_source_id = next_operator_id();
1363
10
            op = std::make_shared<CacheSourceOperatorX>(pool, cache_node_id, cache_source_id,
1364
10
                                                        _params.fragment.query_cache_param);
1365
10
            RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1366
1367
10
            const auto downstream_pipeline_id = cur_pipe->id();
1368
10
            if (!_dag.contains(downstream_pipeline_id)) {
1369
10
                _dag.insert({downstream_pipeline_id, {}});
1370
10
            }
1371
10
            new_pipe = add_pipeline(cur_pipe);
1372
10
            _dag[downstream_pipeline_id].push_back(new_pipe->id());
1373
1374
10
            DataSinkOperatorPtr cache_sink(new CacheSinkOperatorX(
1375
10
                    next_sink_operator_id(), op->node_id(), op->operator_id()));
1376
10
            RETURN_IF_ERROR(new_pipe->set_sink(cache_sink));
1377
10
            return Status::OK();
1378
10
        };
1379
156k
        const bool group_by_limit_opt =
1380
156k
                tnode.agg_node.__isset.agg_sort_info_by_group_key && tnode.limit > 0;
1381
1382
        /// PartitionedAggSourceOperatorX does not support "group by limit opt(#29641)" yet.
1383
        /// If `group_by_limit_opt` is true, then it might not need to spill at all.
1384
156k
        const bool enable_spill = _runtime_state->enable_spill() &&
1385
156k
                                  !tnode.agg_node.grouping_exprs.empty() && !group_by_limit_opt;
1386
156k
        const bool is_streaming_agg = tnode.agg_node.__isset.use_streaming_preaggregation &&
1387
156k
                                      tnode.agg_node.use_streaming_preaggregation &&
1388
156k
                                      !tnode.agg_node.grouping_exprs.empty();
1389
        // TODO: distinct streaming agg does not support spill.
1390
156k
        const bool can_use_distinct_streaming_agg =
1391
156k
                (!enable_spill || is_streaming_agg) && tnode.agg_node.aggregate_functions.empty() &&
1392
156k
                !tnode.agg_node.__isset.agg_sort_info_by_group_key &&
1393
156k
                _params.query_options.__isset.enable_distinct_streaming_aggregation &&
1394
156k
                _params.query_options.enable_distinct_streaming_aggregation;
1395
1396
156k
        if (can_use_distinct_streaming_agg) {
1397
90.7k
            if (need_create_cache_op) {
1398
8
                PipelinePtr new_pipe;
1399
8
                RETURN_IF_ERROR(create_query_cache_operator(new_pipe));
1400
1401
8
                cache_op = op;
1402
8
                op = std::make_shared<DistinctStreamingAggOperatorX>(pool, next_operator_id(),
1403
8
                                                                     tnode, descs);
1404
8
                RETURN_IF_ERROR(new_pipe->add_operator(op, _parallel_instances));
1405
8
                RETURN_IF_ERROR(cur_pipe->operators().front()->set_child(op));
1406
8
                cur_pipe = new_pipe;
1407
90.7k
            } else {
1408
90.7k
                op = std::make_shared<DistinctStreamingAggOperatorX>(pool, next_operator_id(),
1409
90.7k
                                                                     tnode, descs);
1410
90.7k
                RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1411
90.7k
            }
1412
90.7k
        } else if (is_streaming_agg) {
1413
3.07k
            if (need_create_cache_op) {
1414
0
                PipelinePtr new_pipe;
1415
0
                RETURN_IF_ERROR(create_query_cache_operator(new_pipe));
1416
0
                cache_op = op;
1417
0
                op = std::make_shared<StreamingAggOperatorX>(pool, next_operator_id(), tnode,
1418
0
                                                             descs);
1419
0
                RETURN_IF_ERROR(cur_pipe->operators().front()->set_child(op));
1420
0
                RETURN_IF_ERROR(new_pipe->add_operator(op, _parallel_instances));
1421
0
                cur_pipe = new_pipe;
1422
3.07k
            } else {
1423
3.07k
                op = std::make_shared<StreamingAggOperatorX>(pool, next_operator_id(), tnode,
1424
3.07k
                                                             descs);
1425
3.07k
                RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1426
3.07k
            }
1427
63.0k
        } else {
1428
            // create new pipeline to add query cache operator
1429
63.0k
            PipelinePtr new_pipe;
1430
63.0k
            if (need_create_cache_op) {
1431
2
                RETURN_IF_ERROR(create_query_cache_operator(new_pipe));
1432
2
                cache_op = op;
1433
2
            }
1434
1435
63.0k
            if (enable_spill) {
1436
140
                op = std::make_shared<PartitionedAggSourceOperatorX>(pool, tnode,
1437
140
                                                                     next_operator_id(), descs);
1438
62.9k
            } else {
1439
62.9k
                op = std::make_shared<AggSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1440
62.9k
            }
1441
63.0k
            if (need_create_cache_op) {
1442
2
                RETURN_IF_ERROR(cur_pipe->operators().front()->set_child(op));
1443
2
                RETURN_IF_ERROR(new_pipe->add_operator(op, _parallel_instances));
1444
2
                cur_pipe = new_pipe;
1445
63.0k
            } else {
1446
63.0k
                RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1447
63.0k
            }
1448
1449
63.0k
            const auto downstream_pipeline_id = cur_pipe->id();
1450
63.0k
            if (!_dag.contains(downstream_pipeline_id)) {
1451
60.4k
                _dag.insert({downstream_pipeline_id, {}});
1452
60.4k
            }
1453
63.0k
            cur_pipe = add_pipeline(cur_pipe);
1454
63.0k
            _dag[downstream_pipeline_id].push_back(cur_pipe->id());
1455
1456
63.0k
            if (enable_spill) {
1457
140
                sink_ops.push_back(std::make_shared<PartitionedAggSinkOperatorX>(
1458
140
                        pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1459
62.9k
            } else {
1460
62.9k
                sink_ops.push_back(std::make_shared<AggSinkOperatorX>(
1461
62.9k
                        pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1462
62.9k
            }
1463
63.0k
            RETURN_IF_ERROR(cur_pipe->set_sink(sink_ops.back()));
1464
63.0k
            RETURN_IF_ERROR(cur_pipe->sink()->init(tnode, _runtime_state.get()));
1465
63.0k
        }
1466
156k
        break;
1467
156k
    }
1468
156k
    case TPlanNodeType::HASH_JOIN_NODE: {
1469
9.68k
        const auto is_broadcast_join = tnode.hash_join_node.__isset.is_broadcast_join &&
1470
9.68k
                                       tnode.hash_join_node.is_broadcast_join;
1471
9.68k
        const auto enable_spill = _runtime_state->enable_spill();
1472
9.68k
        if (enable_spill && !is_broadcast_join) {
1473
0
            auto tnode_ = tnode;
1474
0
            tnode_.runtime_filters.clear();
1475
0
            auto inner_probe_operator =
1476
0
                    std::make_shared<HashJoinProbeOperatorX>(pool, tnode_, 0, descs);
1477
1478
            // probe side inner sink operator is used to build hash table on probe side when data is spilled.
1479
            // So here use `tnode_` which has no runtime filters.
1480
0
            auto probe_side_inner_sink_operator =
1481
0
                    std::make_shared<HashJoinBuildSinkOperatorX>(pool, 0, 0, tnode_, descs);
1482
1483
0
            RETURN_IF_ERROR(inner_probe_operator->init(tnode_, _runtime_state.get()));
1484
0
            RETURN_IF_ERROR(probe_side_inner_sink_operator->init(tnode_, _runtime_state.get()));
1485
1486
0
            auto probe_operator = std::make_shared<PartitionedHashJoinProbeOperatorX>(
1487
0
                    pool, tnode_, next_operator_id(), descs);
1488
0
            probe_operator->set_inner_operators(probe_side_inner_sink_operator,
1489
0
                                                inner_probe_operator);
1490
0
            op = std::move(probe_operator);
1491
0
            RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1492
1493
0
            const auto downstream_pipeline_id = cur_pipe->id();
1494
0
            if (!_dag.contains(downstream_pipeline_id)) {
1495
0
                _dag.insert({downstream_pipeline_id, {}});
1496
0
            }
1497
0
            PipelinePtr build_side_pipe = add_pipeline(cur_pipe);
1498
0
            _dag[downstream_pipeline_id].push_back(build_side_pipe->id());
1499
1500
0
            auto inner_sink_operator =
1501
0
                    std::make_shared<HashJoinBuildSinkOperatorX>(pool, 0, 0, tnode, descs);
1502
0
            auto sink_operator = std::make_shared<PartitionedHashJoinSinkOperatorX>(
1503
0
                    pool, next_sink_operator_id(), op->operator_id(), tnode_, descs);
1504
0
            RETURN_IF_ERROR(inner_sink_operator->init(tnode, _runtime_state.get()));
1505
1506
0
            sink_operator->set_inner_operators(inner_sink_operator, inner_probe_operator);
1507
0
            sink_ops.push_back(std::move(sink_operator));
1508
0
            RETURN_IF_ERROR(build_side_pipe->set_sink(sink_ops.back()));
1509
0
            RETURN_IF_ERROR(build_side_pipe->sink()->init(tnode_, _runtime_state.get()));
1510
1511
0
            _pipeline_parent_map.push(op->node_id(), cur_pipe);
1512
0
            _pipeline_parent_map.push(op->node_id(), build_side_pipe);
1513
9.68k
        } else {
1514
9.68k
            op = std::make_shared<HashJoinProbeOperatorX>(pool, tnode, next_operator_id(), descs);
1515
9.68k
            RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1516
1517
9.68k
            const auto downstream_pipeline_id = cur_pipe->id();
1518
9.68k
            if (!_dag.contains(downstream_pipeline_id)) {
1519
7.99k
                _dag.insert({downstream_pipeline_id, {}});
1520
7.99k
            }
1521
9.68k
            PipelinePtr build_side_pipe = add_pipeline(cur_pipe);
1522
9.68k
            _dag[downstream_pipeline_id].push_back(build_side_pipe->id());
1523
1524
9.68k
            sink_ops.push_back(std::make_shared<HashJoinBuildSinkOperatorX>(
1525
9.68k
                    pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1526
9.68k
            RETURN_IF_ERROR(build_side_pipe->set_sink(sink_ops.back()));
1527
9.68k
            RETURN_IF_ERROR(build_side_pipe->sink()->init(tnode, _runtime_state.get()));
1528
1529
9.68k
            _pipeline_parent_map.push(op->node_id(), cur_pipe);
1530
9.68k
            _pipeline_parent_map.push(op->node_id(), build_side_pipe);
1531
9.68k
        }
1532
9.68k
        if (is_broadcast_join && _runtime_state->enable_share_hash_table_for_broadcast_join()) {
1533
5.00k
            std::shared_ptr<HashJoinSharedState> shared_state =
1534
5.00k
                    HashJoinSharedState::create_shared(_num_instances);
1535
25.6k
            for (int i = 0; i < _num_instances; i++) {
1536
20.6k
                auto sink_dep = std::make_shared<Dependency>(op->operator_id(), op->node_id(),
1537
20.6k
                                                             "HASH_JOIN_BUILD_DEPENDENCY");
1538
20.6k
                sink_dep->set_shared_state(shared_state.get());
1539
20.6k
                shared_state->sink_deps.push_back(sink_dep);
1540
20.6k
            }
1541
5.00k
            shared_state->create_source_dependencies(_num_instances, op->operator_id(),
1542
5.00k
                                                     op->node_id(), "HASH_JOIN_PROBE");
1543
5.00k
            _op_id_to_shared_state.insert(
1544
5.00k
                    {op->operator_id(), {shared_state, shared_state->sink_deps}});
1545
5.00k
        }
1546
9.68k
        break;
1547
9.68k
    }
1548
4.82k
    case TPlanNodeType::CROSS_JOIN_NODE: {
1549
4.82k
        op = std::make_shared<NestedLoopJoinProbeOperatorX>(pool, tnode, next_operator_id(), descs);
1550
4.82k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1551
1552
4.82k
        const auto downstream_pipeline_id = cur_pipe->id();
1553
4.82k
        if (!_dag.contains(downstream_pipeline_id)) {
1554
4.59k
            _dag.insert({downstream_pipeline_id, {}});
1555
4.59k
        }
1556
4.82k
        PipelinePtr build_side_pipe = add_pipeline(cur_pipe);
1557
4.82k
        _dag[downstream_pipeline_id].push_back(build_side_pipe->id());
1558
1559
4.82k
        sink_ops.push_back(std::make_shared<NestedLoopJoinBuildSinkOperatorX>(
1560
4.82k
                pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1561
4.82k
        RETURN_IF_ERROR(build_side_pipe->set_sink(sink_ops.back()));
1562
4.82k
        RETURN_IF_ERROR(build_side_pipe->sink()->init(tnode, _runtime_state.get()));
1563
4.82k
        _pipeline_parent_map.push(op->node_id(), cur_pipe);
1564
4.82k
        _pipeline_parent_map.push(op->node_id(), build_side_pipe);
1565
4.82k
        break;
1566
4.82k
    }
1567
52.3k
    case TPlanNodeType::UNION_NODE: {
1568
52.3k
        int child_count = tnode.num_children;
1569
52.3k
        op = std::make_shared<UnionSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1570
52.3k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1571
1572
52.3k
        const auto downstream_pipeline_id = cur_pipe->id();
1573
52.3k
        if (!_dag.contains(downstream_pipeline_id)) {
1574
51.7k
            _dag.insert({downstream_pipeline_id, {}});
1575
51.7k
        }
1576
53.9k
        for (int i = 0; i < child_count; i++) {
1577
1.68k
            PipelinePtr build_side_pipe = add_pipeline(cur_pipe);
1578
1.68k
            _dag[downstream_pipeline_id].push_back(build_side_pipe->id());
1579
1.68k
            sink_ops.push_back(std::make_shared<UnionSinkOperatorX>(
1580
1.68k
                    i, next_sink_operator_id(), op->operator_id(), pool, tnode, descs));
1581
1.68k
            RETURN_IF_ERROR(build_side_pipe->set_sink(sink_ops.back()));
1582
1.68k
            RETURN_IF_ERROR(build_side_pipe->sink()->init(tnode, _runtime_state.get()));
1583
            // preset children pipelines. if any pipeline found this as its father, will use the prepared pipeline to build.
1584
1.68k
            _pipeline_parent_map.push(op->node_id(), build_side_pipe);
1585
1.68k
        }
1586
52.3k
        break;
1587
52.3k
    }
1588
52.3k
    case TPlanNodeType::SORT_NODE: {
1589
43.1k
        const auto should_spill = _runtime_state->enable_spill() &&
1590
43.1k
                                  tnode.sort_node.algorithm == TSortAlgorithm::FULL_SORT;
1591
43.1k
        const bool use_local_merge =
1592
43.1k
                tnode.sort_node.__isset.use_local_merge && tnode.sort_node.use_local_merge;
1593
43.1k
        if (should_spill) {
1594
9
            op = std::make_shared<SpillSortSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1595
43.1k
        } else if (use_local_merge) {
1596
40.8k
            op = std::make_shared<LocalMergeSortSourceOperatorX>(pool, tnode, next_operator_id(),
1597
40.8k
                                                                 descs);
1598
40.8k
        } else {
1599
2.31k
            op = std::make_shared<SortSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1600
2.31k
        }
1601
43.1k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1602
1603
43.1k
        const auto downstream_pipeline_id = cur_pipe->id();
1604
43.1k
        if (!_dag.contains(downstream_pipeline_id)) {
1605
43.0k
            _dag.insert({downstream_pipeline_id, {}});
1606
43.0k
        }
1607
43.1k
        cur_pipe = add_pipeline(cur_pipe);
1608
43.1k
        _dag[downstream_pipeline_id].push_back(cur_pipe->id());
1609
1610
43.1k
        if (should_spill) {
1611
9
            sink_ops.push_back(std::make_shared<SpillSortSinkOperatorX>(
1612
9
                    pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1613
43.1k
        } else {
1614
43.1k
            sink_ops.push_back(std::make_shared<SortSinkOperatorX>(
1615
43.1k
                    pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1616
43.1k
        }
1617
43.1k
        RETURN_IF_ERROR(cur_pipe->set_sink(sink_ops.back()));
1618
43.1k
        RETURN_IF_ERROR(cur_pipe->sink()->init(tnode, _runtime_state.get()));
1619
43.1k
        break;
1620
43.1k
    }
1621
43.1k
    case TPlanNodeType::PARTITION_SORT_NODE: {
1622
62
        op = std::make_shared<PartitionSortSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1623
62
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1624
1625
62
        const auto downstream_pipeline_id = cur_pipe->id();
1626
62
        if (!_dag.contains(downstream_pipeline_id)) {
1627
62
            _dag.insert({downstream_pipeline_id, {}});
1628
62
        }
1629
62
        cur_pipe = add_pipeline(cur_pipe);
1630
62
        _dag[downstream_pipeline_id].push_back(cur_pipe->id());
1631
1632
62
        sink_ops.push_back(std::make_shared<PartitionSortSinkOperatorX>(
1633
62
                pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1634
62
        RETURN_IF_ERROR(cur_pipe->set_sink(sink_ops.back()));
1635
62
        RETURN_IF_ERROR(cur_pipe->sink()->init(tnode, _runtime_state.get()));
1636
62
        break;
1637
62
    }
1638
1.64k
    case TPlanNodeType::ANALYTIC_EVAL_NODE: {
1639
1.64k
        op = std::make_shared<AnalyticSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1640
1.64k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1641
1642
1.64k
        const auto downstream_pipeline_id = cur_pipe->id();
1643
1.64k
        if (!_dag.contains(downstream_pipeline_id)) {
1644
1.63k
            _dag.insert({downstream_pipeline_id, {}});
1645
1.63k
        }
1646
1.64k
        cur_pipe = add_pipeline(cur_pipe);
1647
1.64k
        _dag[downstream_pipeline_id].push_back(cur_pipe->id());
1648
1649
1.64k
        sink_ops.push_back(std::make_shared<AnalyticSinkOperatorX>(
1650
1.64k
                pool, next_sink_operator_id(), op->operator_id(), tnode, descs));
1651
1.64k
        RETURN_IF_ERROR(cur_pipe->set_sink(sink_ops.back()));
1652
1.64k
        RETURN_IF_ERROR(cur_pipe->sink()->init(tnode, _runtime_state.get()));
1653
1.64k
        break;
1654
1.64k
    }
1655
1.64k
    case TPlanNodeType::MATERIALIZATION_NODE: {
1656
1.59k
        op = std::make_shared<MaterializationOperator>(pool, tnode, next_operator_id(), descs);
1657
1.59k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1658
1.59k
        break;
1659
1.59k
    }
1660
1.59k
    case TPlanNodeType::INTERSECT_NODE: {
1661
114
        RETURN_IF_ERROR(_build_operators_for_set_operation_node<true>(pool, tnode, descs, op,
1662
114
                                                                      cur_pipe, sink_ops));
1663
114
        break;
1664
114
    }
1665
125
    case TPlanNodeType::EXCEPT_NODE: {
1666
125
        RETURN_IF_ERROR(_build_operators_for_set_operation_node<false>(pool, tnode, descs, op,
1667
125
                                                                       cur_pipe, sink_ops));
1668
125
        break;
1669
125
    }
1670
296
    case TPlanNodeType::REPEAT_NODE: {
1671
296
        op = std::make_shared<RepeatOperatorX>(pool, tnode, next_operator_id(), descs);
1672
296
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1673
296
        break;
1674
296
    }
1675
913
    case TPlanNodeType::TABLE_FUNCTION_NODE: {
1676
913
        op = std::make_shared<TableFunctionOperatorX>(pool, tnode, next_operator_id(), descs);
1677
913
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1678
913
        break;
1679
913
    }
1680
913
    case TPlanNodeType::ASSERT_NUM_ROWS_NODE: {
1681
218
        op = std::make_shared<AssertNumRowsOperatorX>(pool, tnode, next_operator_id(), descs);
1682
218
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1683
218
        break;
1684
218
    }
1685
1.60k
    case TPlanNodeType::EMPTY_SET_NODE: {
1686
1.60k
        op = std::make_shared<EmptySetSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1687
1.60k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1688
1.60k
        break;
1689
1.60k
    }
1690
1.60k
    case TPlanNodeType::DATA_GEN_SCAN_NODE: {
1691
452
        op = std::make_shared<DataGenSourceOperatorX>(pool, tnode, next_operator_id(), descs);
1692
452
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1693
452
        fe_with_old_version = !tnode.__isset.is_serial_operator;
1694
452
        break;
1695
452
    }
1696
2.31k
    case TPlanNodeType::SCHEMA_SCAN_NODE: {
1697
2.31k
        op = std::make_shared<SchemaScanOperatorX>(pool, tnode, next_operator_id(), descs);
1698
2.31k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1699
2.31k
        break;
1700
2.31k
    }
1701
5.90k
    case TPlanNodeType::META_SCAN_NODE: {
1702
5.90k
        op = std::make_shared<MetaScanOperatorX>(pool, tnode, next_operator_id(), descs);
1703
5.90k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1704
5.90k
        break;
1705
5.90k
    }
1706
5.90k
    case TPlanNodeType::SELECT_NODE: {
1707
1.92k
        op = std::make_shared<SelectOperatorX>(pool, tnode, next_operator_id(), descs);
1708
1.92k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1709
1.92k
        break;
1710
1.92k
    }
1711
1.92k
    case TPlanNodeType::REC_CTE_NODE: {
1712
151
        op = std::make_shared<RecCTESourceOperatorX>(pool, tnode, next_operator_id(), descs);
1713
151
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1714
1715
151
        const auto downstream_pipeline_id = cur_pipe->id();
1716
151
        if (!_dag.contains(downstream_pipeline_id)) {
1717
148
            _dag.insert({downstream_pipeline_id, {}});
1718
148
        }
1719
1720
151
        PipelinePtr anchor_side_pipe = add_pipeline(cur_pipe);
1721
151
        _dag[downstream_pipeline_id].push_back(anchor_side_pipe->id());
1722
1723
151
        DataSinkOperatorPtr anchor_sink;
1724
151
        anchor_sink = std::make_shared<RecCTEAnchorSinkOperatorX>(next_sink_operator_id(),
1725
151
                                                                  op->operator_id(), tnode, descs);
1726
151
        RETURN_IF_ERROR(anchor_side_pipe->set_sink(anchor_sink));
1727
151
        RETURN_IF_ERROR(anchor_side_pipe->sink()->init(tnode, _runtime_state.get()));
1728
151
        _pipeline_parent_map.push(op->node_id(), anchor_side_pipe);
1729
1730
151
        PipelinePtr rec_side_pipe = add_pipeline(cur_pipe);
1731
151
        _dag[downstream_pipeline_id].push_back(rec_side_pipe->id());
1732
1733
151
        DataSinkOperatorPtr rec_sink;
1734
151
        rec_sink = std::make_shared<RecCTESinkOperatorX>(next_sink_operator_id(), op->operator_id(),
1735
151
                                                         tnode, descs);
1736
151
        RETURN_IF_ERROR(rec_side_pipe->set_sink(rec_sink));
1737
151
        RETURN_IF_ERROR(rec_side_pipe->sink()->init(tnode, _runtime_state.get()));
1738
151
        _pipeline_parent_map.push(op->node_id(), rec_side_pipe);
1739
1740
151
        break;
1741
151
    }
1742
1.95k
    case TPlanNodeType::REC_CTE_SCAN_NODE: {
1743
1.95k
        op = std::make_shared<RecCTEScanOperatorX>(pool, tnode, next_operator_id(), descs);
1744
1.95k
        RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1745
1.95k
        break;
1746
1.95k
    }
1747
1.95k
    default:
1748
0
        return Status::InternalError("Unsupported exec type in pipeline: {}",
1749
0
                                     print_plan_node_type(tnode.node_type));
1750
667k
    }
1751
665k
    if (_params.__isset.parallel_instances && fe_with_old_version) {
1752
0
        cur_pipe->set_num_tasks(_params.parallel_instances);
1753
0
        op->set_serial_operator();
1754
0
    }
1755
1756
665k
    return Status::OK();
1757
667k
}
1758
// NOLINTEND(readability-function-cognitive-complexity)
1759
// NOLINTEND(readability-function-size)
1760
1761
template <bool is_intersect>
1762
Status PipelineFragmentContext::_build_operators_for_set_operation_node(
1763
        ObjectPool* pool, const TPlanNode& tnode, const DescriptorTbl& descs, OperatorPtr& op,
1764
239
        PipelinePtr& cur_pipe, std::vector<DataSinkOperatorPtr>& sink_ops) {
1765
239
    op.reset(new SetSourceOperatorX<is_intersect>(pool, tnode, next_operator_id(), descs));
1766
239
    RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1767
1768
239
    const auto downstream_pipeline_id = cur_pipe->id();
1769
239
    if (!_dag.contains(downstream_pipeline_id)) {
1770
222
        _dag.insert({downstream_pipeline_id, {}});
1771
222
    }
1772
1773
810
    for (int child_id = 0; child_id < tnode.num_children; child_id++) {
1774
571
        PipelinePtr probe_side_pipe = add_pipeline(cur_pipe);
1775
571
        _dag[downstream_pipeline_id].push_back(probe_side_pipe->id());
1776
1777
571
        if (child_id == 0) {
1778
239
            sink_ops.push_back(std::make_shared<SetSinkOperatorX<is_intersect>>(
1779
239
                    child_id, next_sink_operator_id(), op->operator_id(), pool, tnode, descs));
1780
332
        } else {
1781
332
            sink_ops.push_back(std::make_shared<SetProbeSinkOperatorX<is_intersect>>(
1782
332
                    child_id, next_sink_operator_id(), op->operator_id(), pool, tnode, descs));
1783
332
        }
1784
571
        RETURN_IF_ERROR(probe_side_pipe->set_sink(sink_ops.back()));
1785
571
        RETURN_IF_ERROR(probe_side_pipe->sink()->init(tnode, _runtime_state.get()));
1786
        // prepare children pipelines. if any pipeline found this as its father, will use the prepared pipeline to build.
1787
571
        _pipeline_parent_map.push(op->node_id(), probe_side_pipe);
1788
571
    }
1789
1790
239
    return Status::OK();
1791
239
}
_ZN5doris23PipelineFragmentContext39_build_operators_for_set_operation_nodeILb1EEENS_6StatusEPNS_10ObjectPoolERKNS_9TPlanNodeERKNS_13DescriptorTblERSt10shared_ptrINS_13OperatorXBaseEERSB_INS_8PipelineEERSt6vectorISB_INS_21DataSinkOperatorXBaseEESaISK_EE
Line
Count
Source
1764
114
        PipelinePtr& cur_pipe, std::vector<DataSinkOperatorPtr>& sink_ops) {
1765
114
    op.reset(new SetSourceOperatorX<is_intersect>(pool, tnode, next_operator_id(), descs));
1766
114
    RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1767
1768
114
    const auto downstream_pipeline_id = cur_pipe->id();
1769
114
    if (!_dag.contains(downstream_pipeline_id)) {
1770
105
        _dag.insert({downstream_pipeline_id, {}});
1771
105
    }
1772
1773
420
    for (int child_id = 0; child_id < tnode.num_children; child_id++) {
1774
306
        PipelinePtr probe_side_pipe = add_pipeline(cur_pipe);
1775
306
        _dag[downstream_pipeline_id].push_back(probe_side_pipe->id());
1776
1777
306
        if (child_id == 0) {
1778
114
            sink_ops.push_back(std::make_shared<SetSinkOperatorX<is_intersect>>(
1779
114
                    child_id, next_sink_operator_id(), op->operator_id(), pool, tnode, descs));
1780
192
        } else {
1781
192
            sink_ops.push_back(std::make_shared<SetProbeSinkOperatorX<is_intersect>>(
1782
192
                    child_id, next_sink_operator_id(), op->operator_id(), pool, tnode, descs));
1783
192
        }
1784
306
        RETURN_IF_ERROR(probe_side_pipe->set_sink(sink_ops.back()));
1785
306
        RETURN_IF_ERROR(probe_side_pipe->sink()->init(tnode, _runtime_state.get()));
1786
        // prepare children pipelines. if any pipeline found this as its father, will use the prepared pipeline to build.
1787
306
        _pipeline_parent_map.push(op->node_id(), probe_side_pipe);
1788
306
    }
1789
1790
114
    return Status::OK();
1791
114
}
_ZN5doris23PipelineFragmentContext39_build_operators_for_set_operation_nodeILb0EEENS_6StatusEPNS_10ObjectPoolERKNS_9TPlanNodeERKNS_13DescriptorTblERSt10shared_ptrINS_13OperatorXBaseEERSB_INS_8PipelineEERSt6vectorISB_INS_21DataSinkOperatorXBaseEESaISK_EE
Line
Count
Source
1764
125
        PipelinePtr& cur_pipe, std::vector<DataSinkOperatorPtr>& sink_ops) {
1765
125
    op.reset(new SetSourceOperatorX<is_intersect>(pool, tnode, next_operator_id(), descs));
1766
125
    RETURN_IF_ERROR(cur_pipe->add_operator(op, _parallel_instances));
1767
1768
125
    const auto downstream_pipeline_id = cur_pipe->id();
1769
125
    if (!_dag.contains(downstream_pipeline_id)) {
1770
117
        _dag.insert({downstream_pipeline_id, {}});
1771
117
    }
1772
1773
390
    for (int child_id = 0; child_id < tnode.num_children; child_id++) {
1774
265
        PipelinePtr probe_side_pipe = add_pipeline(cur_pipe);
1775
265
        _dag[downstream_pipeline_id].push_back(probe_side_pipe->id());
1776
1777
265
        if (child_id == 0) {
1778
125
            sink_ops.push_back(std::make_shared<SetSinkOperatorX<is_intersect>>(
1779
125
                    child_id, next_sink_operator_id(), op->operator_id(), pool, tnode, descs));
1780
140
        } else {
1781
140
            sink_ops.push_back(std::make_shared<SetProbeSinkOperatorX<is_intersect>>(
1782
140
                    child_id, next_sink_operator_id(), op->operator_id(), pool, tnode, descs));
1783
140
        }
1784
265
        RETURN_IF_ERROR(probe_side_pipe->set_sink(sink_ops.back()));
1785
265
        RETURN_IF_ERROR(probe_side_pipe->sink()->init(tnode, _runtime_state.get()));
1786
        // prepare children pipelines. if any pipeline found this as its father, will use the prepared pipeline to build.
1787
265
        _pipeline_parent_map.push(op->node_id(), probe_side_pipe);
1788
265
    }
1789
1790
125
    return Status::OK();
1791
125
}
1792
1793
428k
Status PipelineFragmentContext::submit() {
1794
428k
    if (_submitted) {
1795
0
        return Status::InternalError("submitted");
1796
0
    }
1797
428k
    _submitted = true;
1798
1799
428k
    int submit_tasks = 0;
1800
428k
    Status st;
1801
428k
    auto* scheduler = _query_ctx->get_pipe_exec_scheduler();
1802
1.12M
    for (auto& task : _tasks) {
1803
1.92M
        for (auto& t : task) {
1804
1.92M
            st = scheduler->submit(t.first);
1805
1.92M
            DBUG_EXECUTE_IF("PipelineFragmentContext.submit.failed",
1806
1.92M
                            { st = Status::Aborted("PipelineFragmentContext.submit.failed"); });
1807
1.92M
            if (!st) {
1808
0
                cancel(Status::InternalError("submit context to executor fail"));
1809
0
                std::lock_guard<std::mutex> l(_task_mutex);
1810
0
                _total_tasks = submit_tasks;
1811
0
                break;
1812
0
            }
1813
1.92M
            submit_tasks++;
1814
1.92M
        }
1815
1.12M
    }
1816
428k
    if (!st.ok()) {
1817
0
        bool need_remove = false;
1818
0
        {
1819
0
            std::lock_guard<std::mutex> l(_task_mutex);
1820
0
            if (_closed_tasks >= _total_tasks) {
1821
0
                need_remove = _close_fragment_instance();
1822
0
            }
1823
0
        }
1824
        // Call remove_pipeline_context() outside _task_mutex to avoid ABBA deadlock.
1825
0
        if (need_remove) {
1826
0
            _exec_env->fragment_mgr()->remove_pipeline_context({_query_id, _fragment_id});
1827
0
        }
1828
0
        return Status::InternalError("Submit pipeline failed. err = {}, BE: {}", st.to_string(),
1829
0
                                     BackendOptions::get_localhost());
1830
428k
    } else {
1831
428k
        return st;
1832
428k
    }
1833
428k
}
1834
1835
0
void PipelineFragmentContext::print_profile(const std::string& extra_info) {
1836
0
    if (_runtime_state->enable_profile()) {
1837
0
        std::stringstream ss;
1838
0
        for (auto runtime_profile_ptr : _runtime_state->pipeline_id_to_profile()) {
1839
0
            runtime_profile_ptr->pretty_print(&ss);
1840
0
        }
1841
1842
0
        if (_runtime_state->load_channel_profile()) {
1843
0
            _runtime_state->load_channel_profile()->pretty_print(&ss);
1844
0
        }
1845
1846
0
        auto profile_str =
1847
0
                fmt::format("Query {} fragment {} {}, profile, {}", print_id(this->_query_id),
1848
0
                            this->_fragment_id, extra_info, ss.str());
1849
0
        LOG_LONG_STRING(INFO, profile_str);
1850
0
    }
1851
0
}
1852
// If all pipeline tasks binded to the fragment instance are finished, then we could
1853
// close the fragment instance.
1854
// Returns true if the caller should call remove_pipeline_context() **after** releasing
1855
// _task_mutex. We must not call remove_pipeline_context() here because it acquires
1856
// _pipeline_map's shard lock, and this function is called while _task_mutex is held.
1857
// Acquiring _pipeline_map while holding _task_mutex creates an ABBA deadlock with
1858
// dump_pipeline_tasks(), which acquires _pipeline_map first and then _task_mutex
1859
// (via debug_string()).
1860
429k
bool PipelineFragmentContext::_close_fragment_instance() {
1861
429k
    if (_is_fragment_instance_closed) {
1862
0
        return false;
1863
0
    }
1864
429k
    Defer defer_op {[&]() { _is_fragment_instance_closed = true; }};
1865
429k
    _fragment_level_profile->total_time_counter()->update(_fragment_watcher.elapsed_time());
1866
429k
    if (!_need_notify_close) {
1867
426k
        auto st = send_report(true);
1868
426k
        if (!st) {
1869
0
            LOG(WARNING) << fmt::format("Failed to send report for query {}, fragment {}: {}",
1870
0
                                        print_id(_query_id), _fragment_id, st.to_string());
1871
0
        }
1872
426k
    }
1873
    // Print profile content in info log is a tempoeray solution for stream load and external_connector.
1874
    // Since stream load does not have someting like coordinator on FE, so
1875
    // backend can not report profile to FE, ant its profile can not be shown
1876
    // in the same way with other query. So we print the profile content to info log.
1877
1878
429k
    if (_runtime_state->enable_profile() &&
1879
429k
        (_query_ctx->get_query_source() == QuerySource::STREAM_LOAD ||
1880
2.23k
         _query_ctx->get_query_source() == QuerySource::EXTERNAL_CONNECTOR ||
1881
2.23k
         _query_ctx->get_query_source() == QuerySource::GROUP_COMMIT_LOAD)) {
1882
0
        std::stringstream ss;
1883
        // Compute the _local_time_percent before pretty_print the runtime_profile
1884
        // Before add this operation, the print out like that:
1885
        // UNION_NODE (id=0):(Active: 56.720us, non-child: 00.00%)
1886
        // After add the operation, the print out like that:
1887
        // UNION_NODE (id=0):(Active: 56.720us, non-child: 82.53%)
1888
        // We can easily know the exec node execute time without child time consumed.
1889
0
        for (auto runtime_profile_ptr : _runtime_state->pipeline_id_to_profile()) {
1890
0
            runtime_profile_ptr->pretty_print(&ss);
1891
0
        }
1892
1893
0
        if (_runtime_state->load_channel_profile()) {
1894
0
            _runtime_state->load_channel_profile()->pretty_print(&ss);
1895
0
        }
1896
1897
0
        LOG_INFO("Query {} fragment {} profile:\n {}", print_id(_query_id), _fragment_id, ss.str());
1898
0
    }
1899
1900
429k
    if (_query_ctx->enable_profile()) {
1901
2.23k
        _query_ctx->add_fragment_profile(_fragment_id, collect_realtime_profile(),
1902
2.23k
                                         collect_realtime_load_channel_profile());
1903
2.23k
    }
1904
1905
    // Return whether the caller needs to remove from the pipeline map.
1906
    // The caller must do this after releasing _task_mutex.
1907
429k
    return !_need_notify_close;
1908
429k
}
1909
1910
1.91M
void PipelineFragmentContext::decrement_running_task(PipelineId pipeline_id) {
1911
    // If all tasks of this pipeline has been closed, upstream tasks is never needed, and we just make those runnable here
1912
1.91M
    DCHECK(_pip_id_to_pipeline.contains(pipeline_id));
1913
1.91M
    if (_pip_id_to_pipeline[pipeline_id]->close_task()) {
1914
671k
        if (_dag.contains(pipeline_id)) {
1915
353k
            for (auto dep : _dag[pipeline_id]) {
1916
353k
                _pip_id_to_pipeline[dep]->make_all_runnable(pipeline_id);
1917
353k
            }
1918
282k
        }
1919
671k
    }
1920
1.91M
    bool need_remove = false;
1921
1.91M
    {
1922
1.91M
        std::lock_guard<std::mutex> l(_task_mutex);
1923
1.91M
        ++_closed_tasks;
1924
1.91M
        if (_closed_tasks >= _total_tasks) {
1925
429k
            need_remove = _close_fragment_instance();
1926
429k
        }
1927
1.91M
    }
1928
    // Call remove_pipeline_context() outside _task_mutex to avoid ABBA deadlock.
1929
1.91M
    if (need_remove) {
1930
426k
        _exec_env->fragment_mgr()->remove_pipeline_context({_query_id, _fragment_id});
1931
426k
    }
1932
1.91M
}
1933
1934
53.2k
std::string PipelineFragmentContext::get_load_error_url() {
1935
53.2k
    if (const auto& str = _runtime_state->get_error_log_file_path(); !str.empty()) {
1936
0
        return to_load_error_http_path(str);
1937
0
    }
1938
158k
    for (auto& tasks : _tasks) {
1939
264k
        for (auto& task : tasks) {
1940
264k
            if (const auto& str = task.second->get_error_log_file_path(); !str.empty()) {
1941
163
                return to_load_error_http_path(str);
1942
163
            }
1943
264k
        }
1944
158k
    }
1945
53.0k
    return "";
1946
53.2k
}
1947
1948
53.2k
std::string PipelineFragmentContext::get_first_error_msg() {
1949
53.2k
    if (const auto& str = _runtime_state->get_first_error_msg(); !str.empty()) {
1950
0
        return str;
1951
0
    }
1952
158k
    for (auto& tasks : _tasks) {
1953
264k
        for (auto& task : tasks) {
1954
264k
            if (const auto& str = task.second->get_first_error_msg(); !str.empty()) {
1955
163
                return str;
1956
163
            }
1957
264k
        }
1958
158k
    }
1959
53.0k
    return "";
1960
53.2k
}
1961
1962
431k
Status PipelineFragmentContext::send_report(bool done) {
1963
431k
    Status exec_status = _query_ctx->exec_status();
1964
    // If plan is done successfully, but _is_report_success is false,
1965
    // no need to send report.
1966
    // Load will set _is_report_success to true because load wants to know
1967
    // the process.
1968
431k
    if (!_is_report_success && done && exec_status.ok()) {
1969
384k
        return Status::OK();
1970
384k
    }
1971
1972
    // If both _is_report_success and _is_report_on_cancel are false,
1973
    // which means no matter query is success or failed, no report is needed.
1974
    // This may happen when the query limit reached and
1975
    // a internal cancellation being processed
1976
    // When limit is reached the fragment is also cancelled, but _is_report_on_cancel will
1977
    // be set to false, to avoid sending fault report to FE.
1978
46.6k
    if (!_is_report_success && !_is_report_on_cancel) {
1979
344
        if (done) {
1980
            // if done is true, which means the query is finished successfully, we can safely close the fragment instance without sending report to FE, and just return OK status here.
1981
344
            return Status::OK();
1982
344
        }
1983
0
        return Status::NeedSendAgain("");
1984
344
    }
1985
1986
46.3k
    std::vector<RuntimeState*> runtime_states;
1987
1988
116k
    for (auto& tasks : _tasks) {
1989
169k
        for (auto& task : tasks) {
1990
169k
            runtime_states.push_back(task.second.get());
1991
169k
        }
1992
116k
    }
1993
1994
46.3k
    std::string load_eror_url = _query_ctx->get_load_error_url().empty()
1995
46.3k
                                        ? get_load_error_url()
1996
18.4E
                                        : _query_ctx->get_load_error_url();
1997
46.3k
    std::string first_error_msg = _query_ctx->get_first_error_msg().empty()
1998
46.3k
                                          ? get_first_error_msg()
1999
18.4E
                                          : _query_ctx->get_first_error_msg();
2000
2001
46.3k
    ReportStatusRequest req {.status = exec_status,
2002
46.3k
                             .runtime_states = runtime_states,
2003
46.3k
                             .done = done || !exec_status.ok(),
2004
46.3k
                             .coord_addr = _query_ctx->coord_addr,
2005
46.3k
                             .query_id = _query_id,
2006
46.3k
                             .fragment_id = _fragment_id,
2007
46.3k
                             .fragment_instance_id = TUniqueId(),
2008
46.3k
                             .backend_num = -1,
2009
46.3k
                             .runtime_state = _runtime_state.get(),
2010
46.3k
                             .load_error_url = load_eror_url,
2011
46.3k
                             .first_error_msg = first_error_msg,
2012
46.3k
                             .cancel_fn = [this](const Status& reason) { cancel(reason); }};
2013
2014
46.3k
    return _report_status_cb(
2015
46.3k
            req, std::dynamic_pointer_cast<PipelineFragmentContext>(shared_from_this()));
2016
46.6k
}
2017
2018
0
size_t PipelineFragmentContext::get_revocable_size(bool* has_running_task) const {
2019
0
    size_t res = 0;
2020
    // _tasks will be cleared during ~PipelineFragmentContext, so that it's safe
2021
    // here to traverse the vector.
2022
0
    for (const auto& task_instances : _tasks) {
2023
0
        for (const auto& task : task_instances) {
2024
0
            if (task.first->is_running()) {
2025
0
                LOG_EVERY_N(INFO, 50) << "Query: " << print_id(_query_id)
2026
0
                                      << " is running, task: " << (void*)task.first.get()
2027
0
                                      << ", is_running: " << task.first->is_running();
2028
0
                *has_running_task = true;
2029
0
                return 0;
2030
0
            }
2031
2032
0
            size_t revocable_size = task.first->get_revocable_size();
2033
0
            if (revocable_size >= SpillFile::MIN_SPILL_WRITE_BATCH_MEM) {
2034
0
                res += revocable_size;
2035
0
            }
2036
0
        }
2037
0
    }
2038
0
    return res;
2039
0
}
2040
2041
0
std::vector<PipelineTask*> PipelineFragmentContext::get_revocable_tasks() const {
2042
0
    std::vector<PipelineTask*> revocable_tasks;
2043
0
    for (const auto& task_instances : _tasks) {
2044
0
        for (const auto& task : task_instances) {
2045
0
            size_t revocable_size_ = task.first->get_revocable_size();
2046
2047
0
            if (revocable_size_ >= SpillFile::MIN_SPILL_WRITE_BATCH_MEM) {
2048
0
                revocable_tasks.emplace_back(task.first.get());
2049
0
            }
2050
0
        }
2051
0
    }
2052
0
    return revocable_tasks;
2053
0
}
2054
2055
61
std::string PipelineFragmentContext::debug_string() {
2056
61
    std::lock_guard<std::mutex> l(_task_mutex);
2057
61
    fmt::memory_buffer debug_string_buffer;
2058
61
    fmt::format_to(debug_string_buffer,
2059
61
                   "PipelineFragmentContext Info: _closed_tasks={}, _total_tasks={}, "
2060
61
                   "need_notify_close={}, fragment_id={}, _rec_cte_stage={}\n",
2061
61
                   _closed_tasks, _total_tasks, _need_notify_close, _fragment_id, _rec_cte_stage);
2062
350
    for (size_t j = 0; j < _tasks.size(); j++) {
2063
289
        fmt::format_to(debug_string_buffer, "Tasks in instance {}:\n", j);
2064
771
        for (size_t i = 0; i < _tasks[j].size(); i++) {
2065
482
            fmt::format_to(debug_string_buffer, "Task {}: {}\n", i,
2066
482
                           _tasks[j][i].first->debug_string());
2067
482
        }
2068
289
    }
2069
2070
61
    return fmt::to_string(debug_string_buffer);
2071
61
}
2072
2073
std::vector<std::shared_ptr<TRuntimeProfileTree>>
2074
2.23k
PipelineFragmentContext::collect_realtime_profile() const {
2075
2.23k
    std::vector<std::shared_ptr<TRuntimeProfileTree>> res;
2076
2077
    // we do not have mutex to protect pipeline_id_to_profile
2078
    // so we need to make sure this funciton is invoked after fragment context
2079
    // has already been prepared.
2080
2.23k
    if (!_prepared) {
2081
0
        std::string msg =
2082
0
                "Query " + print_id(_query_id) + " collecting profile, but its not prepared";
2083
0
        DCHECK(false) << msg;
2084
0
        LOG_ERROR(msg);
2085
0
        return res;
2086
0
    }
2087
2088
    // Make sure first profile is fragment level profile
2089
2.23k
    auto fragment_profile = std::make_shared<TRuntimeProfileTree>();
2090
2.23k
    _fragment_level_profile->to_thrift(fragment_profile.get(), _runtime_state->profile_level());
2091
2.23k
    res.push_back(fragment_profile);
2092
2093
    // pipeline_id_to_profile is initialized in prepare stage
2094
4.03k
    for (auto pipeline_profile : _runtime_state->pipeline_id_to_profile()) {
2095
4.03k
        auto profile_ptr = std::make_shared<TRuntimeProfileTree>();
2096
4.03k
        pipeline_profile->to_thrift(profile_ptr.get(), _runtime_state->profile_level());
2097
4.03k
        res.push_back(profile_ptr);
2098
4.03k
    }
2099
2100
2.23k
    return res;
2101
2.23k
}
2102
2103
std::shared_ptr<TRuntimeProfileTree>
2104
2.23k
PipelineFragmentContext::collect_realtime_load_channel_profile() const {
2105
    // we do not have mutex to protect pipeline_id_to_profile
2106
    // so we need to make sure this funciton is invoked after fragment context
2107
    // has already been prepared.
2108
2.23k
    if (!_prepared) {
2109
0
        std::string msg =
2110
0
                "Query " + print_id(_query_id) + " collecting profile, but its not prepared";
2111
0
        DCHECK(false) << msg;
2112
0
        LOG_ERROR(msg);
2113
0
        return nullptr;
2114
0
    }
2115
2116
5.07k
    for (const auto& tasks : _tasks) {
2117
10.0k
        for (const auto& task : tasks) {
2118
10.0k
            if (task.second->load_channel_profile() == nullptr) {
2119
0
                continue;
2120
0
            }
2121
2122
10.0k
            auto tmp_load_channel_profile = std::make_shared<TRuntimeProfileTree>();
2123
2124
10.0k
            task.second->load_channel_profile()->to_thrift(tmp_load_channel_profile.get(),
2125
10.0k
                                                           _runtime_state->profile_level());
2126
10.0k
            _runtime_state->load_channel_profile()->update(*tmp_load_channel_profile);
2127
10.0k
        }
2128
5.07k
    }
2129
2130
2.23k
    auto load_channel_profile = std::make_shared<TRuntimeProfileTree>();
2131
2.23k
    _runtime_state->load_channel_profile()->to_thrift(load_channel_profile.get(),
2132
2.23k
                                                      _runtime_state->profile_level());
2133
2.23k
    return load_channel_profile;
2134
2.23k
}
2135
2136
// Collect runtime filter IDs registered by all tasks in this PFC.
2137
// Used during recursive CTE stage transitions to know which filters to deregister
2138
// before creating the new PFC for the next recursion round.
2139
// Called from rerun_fragment(wait_for_destroy) while tasks are still closing.
2140
// Thread safety: safe because _tasks is structurally immutable after prepare() —
2141
// the vector sizes do not change, and individual RuntimeState filter sets are
2142
// written only during open() which has completed by the time we reach rerun.
2143
3.28k
std::set<int> PipelineFragmentContext::get_deregister_runtime_filter() const {
2144
3.28k
    std::set<int> result;
2145
12.4k
    for (const auto& _task : _tasks) {
2146
20.9k
        for (const auto& task : _task) {
2147
20.9k
            auto set = task.first->runtime_state()->get_deregister_runtime_filter();
2148
20.9k
            result.merge(set);
2149
20.9k
        }
2150
12.4k
    }
2151
3.28k
    if (_runtime_state) {
2152
3.28k
        auto set = _runtime_state->get_deregister_runtime_filter();
2153
3.28k
        result.merge(set);
2154
3.28k
    }
2155
3.28k
    return result;
2156
3.28k
}
2157
2158
431k
void PipelineFragmentContext::_release_resource() {
2159
431k
    std::lock_guard<std::mutex> l(_task_mutex);
2160
    // The memory released by the query end is recorded in the query mem tracker.
2161
431k
    SCOPED_SWITCH_THREAD_MEM_TRACKER_LIMITER(_query_ctx->query_mem_tracker());
2162
431k
    auto st = _query_ctx->exec_status();
2163
1.13M
    for (auto& _task : _tasks) {
2164
1.13M
        if (!_task.empty()) {
2165
1.13M
            _call_back(_task.front().first->runtime_state(), &st);
2166
1.13M
        }
2167
1.13M
    }
2168
431k
    _tasks.clear();
2169
431k
    _dag.clear();
2170
431k
    _pip_id_to_pipeline.clear();
2171
431k
    _pipelines.clear();
2172
431k
    _sink.reset();
2173
431k
    _root_op.reset();
2174
431k
    _runtime_filter_mgr_map.clear();
2175
431k
    _op_id_to_shared_state.clear();
2176
431k
}
2177
2178
#include "common/compile_check_end.h"
2179
} // namespace doris